TEXT FLY WITHIN
THE BOOK ONLY
u< OU_1 60543
OTJP— 7RR— 13-6-7r>— 5,000
OSMAN1A UNIVERSITY LIBRARY
Call No. ' Accession No.
Author
Title .
This book should he returned on »r hefoic the date last marked below.
BUSINESS CYCLES
VOLUME II
BUSINESS CYCLES
A Theoretical, Historical,
and Statistical Analysis of the
Capitalist Process
BY
JOSEPH A. SCHUMPETER
Professor of Economics, Harvard University
VOLUME II
FIRST EDITION
McGRAW-HILL BOOK COMPANY, INC.
NEW YORK AND LONDON
1939
COPYRIGHT, 1939, BY THE
McGRAW-HiLL BOOK COMPANY, INC.
PRINTED IN THE UNITED STATES OP AMERICA
All rights reserved. This book, or
parts thereof, may not be reproduced
in any form without permission of
the publishers.
THE MAPLE PRESS COMPANY YORK, PA.
Contents
VOLUME II
CHAPTER VIII
THE PRICE LEVEL 449
A. A Preliminary Warning Concerning the Causal and the Sympto-
matic Importance of Variations in Price Level 449
B. The Theory of the Price Level 452
C. The Practical Question 458
D. Analysis of the Behavior of Price-level Series; the Pulse Charts; the
Gold Factor 461
E. Group Prices; the Salient Fact of Covariation 475
CHAPTER IX
PHYSICAL QUANTITIES. EMPLOYMENT 483
A. Individual and Composite Quantities; Three Methods of Indicating
Variations in Total Output 483
B. The Analysis of the Trend in Total Industrial Output; the Problem
of Retardation 491
C. The Cyclical Behavior of the Physical Volume of Production . . . 500
Behavior of output and price level compared 506
D. Employment of Labor 509
Discussion of the English unemployment series; absorption of cyclical unemployment at
riaing wage rates, cyclical behavior ... 516
CHAPTER X
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 520
A. Prices and Quantities of Individual Commodities (Including
Services) 520
1. Expectation for infinite variety of responses to cyclical situations 521
2. The competitive case; the "nominal" effect of cyclical variations in expenditure . . 522
3. Reaction of competitive industries to "real" changes in receipts 524
4. Illustrative discussion of various types of behavior 525
B. Special Cases 528
Recent work on price analysis 528
The case of coffee 529
The hog cycle and other "animal cycles" 530
C. The Cycle in Shipbuilding; Professor Tinbergen's Model 583
v
vi CONTENTS
PAGB
D. Entrepreneurial Price Policies 535
The case of entrepreneurs confronted by a demand of infinite elasticity 535
The imperfectly competitive situation; oligopoly in particular . 536
Rigidity or stability of price . 538
CHAPTER XI
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 544
A. Some Propositions about Money 544
B. System Expenditure (Outside Clearings); the Cyclical Behavior of
Producers' and Consumers' Expenditure * 548
C. National Income and Wages 561
1. The behavior of national income as to trend and cyclical variations; comparison with
other series; the profit item .... 561
2. Factual and conceptual difficulties about wages, various definitions 564
8. Analysis of trends in, and cyclical variations of, wage bills and wage rates , 567
4. The question of lags ... . . . 572
5. A difficulty in theoretical interpretation. 574
D. Deposits and Loans 578
1. Possible ways of financing an act of expenditure. . ... . 578
2. The relation between (outside) deposits, loans, and clearings . .581
3. Qualification of the expectation for parallelism . 584
4. On underspending, nonborrowing, and the role of saving in the cycle. 587
5. Customers' balances and bank loans in detail; behavior of various monetary ratios . 594
6. Real and monetary investment ... 598
CHAPTER XII
THE RATE OF INTEREST 602
A. Earlier Argument Resumed 602
1. The concepts of a demand for balances and of an equilibrium rate of interest; the
adapted rate ... . . . 602
2. The secondary wave and the cyclical shifts in the demand for balances; the lag of
interest ... . 604
8. The "supply" of balances . . 606
4. The position of interest in the cyclical process; capitalisation of quasirents . . . 607
5. The pricing of titles to future balances 611
B. Discussion of Various Rates 614
The historical course of interest rates 614
1. Factual difficulties . . . 615
2. The mortgage rate ... 617
3. Bond yields 618
4. The rates of the open market 622
5. The rates of the Bank of England and of the Reichsbank 625
C. Discussion of the Time Shape of Interest Rate 626
1. The absence of trend in the series 627
2. The cyclical behavior of interest ... 628
8. Interest and profits, pig iron production, total output, and price level 632
4. The consequential character of interest and Professor von Hayek's theory; the lag in
short and long rates 634
CHAPTER XIII
THE CENTRAL MARKET AND THE STOCK EXCHANGE 639
A. Banks and the Pulse of Industry 639
1. Limitations to the initiative of banks; banks are normally not "loaned up" . 639
2. Member banks' operations in the open market; their secondary reserves, investments,
and loans 643
CONTENTS vii
PAGE
8. Other factors affecting banks' investments; the investments of industrial or promotion
banks 646
B. The Central Market (in an Isolated Domain) 648
Position and influence of bankers' banks; their behavior with respect to cyclical situations 651
C. The Cyclical Aspects of International Relations 666
1. Exports and imports under the influence of cyclical fluctuations; the role of central
banks 666
2. Capital movements and central bank policy; the English case; pivotal importance of
short claims on foreign countries . .... 668
3. The London gold market; the gold policy of the Bank of England; gold stock, bank rate,
and price level 675
D. Stock Exchange Series 678
1. Cyclical nature and cyclical effects of the trade in stocks and bonds; peculiarities of the
pricing of stocks ... . . 679
2. Theory of stock-market " tendencies " ... . . 682
3. Analysis of the behavior of stock prices; various covariations . 684
4. Banks and stock speculation . . . 689
CHAPTER XIV
1919-1929 692
A. Postwar Events and Postwar Problems (Introduction) 692
B. Comments on Postwar Patterns 695
1. Some factors and symptoms of change in the social structure 697
2. The world war and its consequences as external factors 700
3. Survey of postwar foreign policies and economic relations; reasons for the failure of the
provisional arrangements arrived at .... 702
4. Various other types of war effects exemplified . . . . 704
5. Postwar protectionism and the refusal of this country "to accept its creditor position" 705
C. Further Comments on Postwar Conditions in Our Three Countries 708
1. The United States; mentality of the country and its fiscal policy . . 708
2. Digression on the effects of taxation . 710
8. Germany; the postwar situation and the social atmosphere, expenditure and fiscal policy 714
4. Facts and theory of the German "capital imports" . 718
5. England; the postwar situation; monetary policy; the social atmosphere; fiscal policy 722
D. Outlines of Economic History from 1919 to 1929 782
1. Agrarian developments; diagnosis of the agrarian depressions in the United States,
England, and Germany . . 782
2. The building booms of the twenties in the United States, Germany, and England . 743
E. The "Industrial Revolution'* of the Twenties 753
General features ... 753
1. English developments; external factors; motorcars; iron and steel, electricity; the
expanding industries; cycles and phases ... .... 755
2. German developments, "rationalization"; public enterprise, mergers, the "crisis of
concerns"; the potash problem; iron and steel; organizational innovations; the chem-
ical industry; rayon in particular; cycles and phases . . 759
8. Developments in the United States; the general pattern; aviation; power production,
electrical equipment; motorcars, oil and rubber, the chemical industry, rayon and
textiles; iron and steel, aluminum, copper; value added and value added divided by
payroll in industries, whose value added was 50 millions or more in 1929; detailed
discussion of cyclical fluctuations through 1929 785
F. The Behavior of Systematic Series from 1919 to 1929 794
I. Output (a), prices (6), interest rates (c). ... . ... 794
c. 1. Total industrial output and output per employee . . 795
2. Producers' and consumers' goods; equipment goods; "overproduction" and
"excess capacity" . . 799
viii CONTENTS
PAGE
8. Employment and the absorption of unemployment 803
6. Price levels as shaped by the evolutionary process and by reaction to war inflation 807
1. Analysis of variations in the American price level 808
2. The English case 810
8. The German case 811
e. Interest rates 811
1. The behavior of interest rates in the United States 812
2. The German case 815
3. The English case .... 816
II. a. Debits outside New York, national income, and payroll in the United States . . 817
6. Corporate accumulation in the United States . 820
e. Total realized income and consumers' expenditure; households' borrowings; infer-
ence about households' savings ... 823
d. German and British national incomes, wage bills, consumers' expenditures, and
savings 827
e. The behavior of profits in the United States 830
/. The behavior of wage rates 834
1. American wage rates during the twenties 835
2. Their relation to the amount of unemployment and to prosperities .... 838
3. German wage rates during the twenties 844
4. Wage rates in the United Kingdom during the twenties 847
III. Series descriptive of processes in the sphere of money and member bank credit, mainly
in the United States . . 850
a. The nature of time deposits. . . ... , 851
b. Outside deposits and loans and the structural change in assets 857
c. Banks' investments and general business practice 862
d. German and English banking developments . . 868
IV. Stock speculation and the monetary investment process, mainly in the United States 870
a. The theory of brokers' loans ... . . . . 870
6. The speculative outburst; call rate; minor points; issues for real investment. . 874
c. Stock exchange and investment processes in Germany and in England . . . 880
V. Central banking in the United States and England . . 885
a. The American postwar situation . ... . . 885
6. The mechanics of Federal reserve credit . . . . . 888
c. The reserve system's postwar policy; conclusions; its "failure to prevent the
depression" . . . . . . ... 894
d. Comments on the policy of the Bank of England ... . 904
CHAPTER XV
THE WORLD CRISIS AND AFTER 906
A. The World Crisis and the Cyclical Schema 906
B. 1930 911
1. The United States 911
2. England ... 917
3. Germany ... . 920
C. 1931 and 1932 924
1. Physical Production, on the downgrade and at the bottom . . .... 924
2. Incidents, Accidents, and Policy in Germany . ... . . 930
3. Incidents, Accidents, and Policy in the United States 936
4. The behavior of other American and German time series around the bottom of the
depression 944
D. The United Kingdom, 1931-1938 954
1. The abandonment of the Gold Standard 955
2. The subsequent management of money and credit 958
8. Some effects of abandonment and cognate policies . 959
a. The relief to exports 960
b. England's conversion to protectionism and the Ottawa achievement 961
4. England's industrial processes; the great building boom in particular . . . 963
5. The testimony of time series, especially of output, unemployment, prices, wages, in-
come, clearings, issues, stock prices 966
CONTENTS ix
PAQB
E. The State-Directed Economy of Germany, 1933-1938 971
Employment and production 971
Efforts toward autarky 973
Policy in the recovery phase 974
Government expenditure in the prosperity phase . . . . .... 976
Prices, wages, incomes, consumption ... 977
Monetary management and the behavior of monetary and banking series 979
F. Recovery and Recovery Policy in the United States from 1933-1935 983
The problem . .983
1. " Minor " measures surveyed ... . . 986
2. Recovery and the AA A. . . . . . . . 988
3. Recovery and the NRA . . . 992
4. Monetary policy .... ... . . . 996
5. Income-generating expenditure, facts and theory . . . 1001
6. The statistical picture ... . . . 1006
G. The Disappointing Juglar 1011
The problem raised by American developments since 1935 , .1011
1. Time series contours . . 1013
2. The money market and the banking sphere; behavior of price level 1016
8. The industrial process . . . 1020
4. The reciprocal trade agreements, monetary "restriction" and "expansion"; the
stoppage of income generation . . . 1026
5. Stagnation of the capitalist process . . . 1032
a. {The theory of vanishing investment opportunity . . . 1032
b. Its inadequacy 1036
c. The alternative explanation by means of the social atmosphere resulting from cap-
italist evolution 1038
d. Reasons for believing this explanation to be adequate 1044
APPENDIX: DESCRIPTION OF THE STATISTICAL MATERIAL EMBODIED IN
CHARTS I-LX 1051
Index 1079
CHAPTER VIII
The Price Level
A. A Preliminary Warning. — The fact that price-level series are the
first to be discussed should not be interpreted to mean that we consider
them first in either causal or symptomatic importance. Businessmen,
politicians, and many economists unite in drawing a picture which grossly
exaggerates the role of price movements in the cyclical process. While
for obvious reasons there is some excuse for this in the case of businessmen
and politicians, only faulty analysis can account for it in the case of econo-
mists. The very definition, "the crisis is a break in prices," and still
more propositions such as "the collapse of the price system is the real
cause of a depression," betray failure to realize that the cycle is a process
within which all elements of the economic system interact in certain char-
acteristic ways and that no one element can be singled out for the role of
prime mover. The mistake involved is, thus, much more fundamental
than it would be if there really were any sense in searching the system of
economic quantities for any single element responsible for the cycle, and if
the theory alluded to merely seized upon a wrong one.
It should be abundantly clear from our theoretical discussion in the
third and fourth chapters and from our historical discussion in the sixth
and seventh, that price movements are not the all-important factor in the
business cycle that they are sometimes held to be. We cannot too often
repeat that price movements are not causal to the prosperity phases of our
process; that prosperity can, and sometimes actually does, start from a
falling price level; that those innovations which "ignite *' prosperity do not
presuppose, though they induce, an increase in prices but are profitable at
the existing level; and that innovations which are not profitable at the
existing level are maladjustments in the same sense as are all other
operations that pay only because prices are rising. But rising prices,
being part of the mechanism — as far as innovations are financed by credit
creation and under ideal conditions of full employment — by which factors
of production are directed toward their new employments, do create
additional margins of profit. They have a dislocating influence which
directly and also indirectly, by inducing error and condoning incapacity
and misconduct, accounts for much that happens in the subsequent
449
450 BUSINESS CYCLES
periods of liquidation and helps to make them abnormal, i.e., to turn them
into depressions.
Again, a fall in prices is not the same as a fall in money earnings, which
in turn is not the same as a fall in real earnings. It is necessary, in order
to get into a sound frame of mind in matters of the business cycle, to divest
oneself of the prejudice that a general fall in prices, such as would normally
occur by virtue of the working of the cyclical mechanism, is in itself a
catastrophe or necessarily productive of catastrophes," that falling prices
must always spell misery; that they necessarily increase the burden of
debt in any sense in which that would be synonymous with causing
trouble; that they are incompatible with prosperous business; or that they
are simply an unmixed evil which is to be prevented at any cost and which
can be prevented without impairing the efficiency of the capitalist
machine. But the cyclical fall in price level that occurs in recession is an
element in a process of adjustment to the changes wrought by what hap-
pened in prosperity, and that process deals harshly with many people.
As rising prices, so do falling prices become an intermediate cause of
secondary phenomena, and they may also acquire a momentum of their
own and then move, particularly during depression, in a way that does
not lead to adjustment but spells additional disturbance.1
Our analysis, however, leads us to believe that at least the symptom-
atic value of price movements should be great. So it is, of course, but
less so than we might think. For, as we shall see presently and as is
obvious from common experience, neither prices of individual commodi-
ties, influenced as they must be by the particular conditions and policies
prevailing in individual industries, nor the whole world of prices, however
measured, can really be expected to keep a consistent relation to other
series representing industrial conditions or to the processes that lie behind
them. Most price series, to be sure, display traces of the cyclical move-
ment and on the whole the association is fairly satisfactory. But we must
be careful about prediction in any individual case and refrain from draw-
ing far-reaching conclusions about cycles for periods in which price data
are practically all we have.
Series of individual prices give rise to many complicated questions.
Those that refer to finished products are fertile of practically insoluble
difficulties about quality, local differences, and so on. However, these
series have at least an obvious meaning. The same may, with some
1 The distinction between the fall of prices in recession and in depression may facilitate
agreement, at least as to diagnosis if not as to policy. So many extra-economic considera-
tions and so many valuations of the interests affected enter into the latter that even perfect
agreement in economic argument would help but little toward agreement about measures.
Even the purely economic argument cannot be fully presented in this book. Various
contributions to it, however, have been and will be offered at various turns of our way.
THE PRICE LEVEL 451
qualifications, be said of composites of prices of different, but related,
commodities, which we will call Group Prices. Everyone knows the
reasons that prompt us to construct price indices of motorcars in general
or textiles in general. It is also common knowledge that such composites
may be very misleading. We shall not go into the question of principle
involved, the roots of which stretch far into general theory, but will
confine ourselves to one remark. An index of this sort may give a picture
that is free from many idiosyncracies of the price movements of the indi-
vidual commodities which enter into it and may be useful for many purposes.
It is, however, inadequate for our purpose because the internal shifts that
such an index blots out may be the essential thing. If innovation results
in one of the commodities in such a group pushing out another, the varia-
tions of their prices relative to each other are fundamental to the under-
standing of the cycle or cycles during which they occur. Obviously a
composite consisting of the prices of cotton, wool, silk, and linen textiles
from 1780 to 1830 would be almost valueless in a study of the business
cycles of that period.1
But the general level of prices raises an entirely different problem.
Here it is the very meaning of the thing that becomes doubtful. It has
even been doubted whether there is any meaning to it at all or, less
radically, whether the general level of prices measures a definite some-
thing that exists as such or is merely a statistical figure measuring, for
instance, the common movement of individual prices or — not necessarily
the same thing — what is common to the movements of all individual
prices.2 Most statisticians and economists in the long array of authors
from Dutot and Carli onward did not bother about this at all and went
ahead on such common-sense considerations as plausibility or absurdity of
results and convenience of calculation, taking for granted the economic
meaning of the procedure. It is hardly uncharitable to say that, with
1 It should be observed, moreover, that indices of group prices will occasionally blur
the cyclical picture in other ways also. We need only imagine a case in which all the con-
stituent prices display strong cycles identical in everything but phase. There is plenty of
justification for the toning down of the fluctuations that will result. But the specifically
cyclical aspect of the prices will be partly, in the limiting case entirely, lost. This should
be borne in mind also in the case of the general price level, which of course will display a
smaller amplitude of fluctuations if, as is unavoidably the case, constituents do not move
exactly in step. This is as it should be and will not mislead, provided we confine ourselves
to considering the general price level as a monetary parameter only. But if we took it to
measure the average amplitude of fluctuations in individual prices, we should go completely
astray.
2 Businessmen and representatives of business interests, when they speak of price levels,
mostly mean only the price of their own product. In a pamphlet on the desirability of
stabilization of the price level the writer found that the expression was used 86 times, with-
out meaning once what it ought to mean, or in fact anything that can be usefully designated
by the term.
452 BUSINESS CYCLES
the outstanding exceptions of .Jevons and Edgeworth, they evolved and
applied methods of measurement without knowing exactly what it was
they wanted to measure. Progress has been made nevertheless, par-
ticularly by systematizing and analyzing criteria of choice between various
formulae — a progress chiefly associated with the names of Irving Fisher
(The Making of Index Numbers, 1922) and L. von Bortkiewicz (see
Geld, II. Die Messung des Geldwertes; Handworterbuch der Staats-
wissenschaften) . And other lines of advance have been* opened up by
Pigou, Haberler, Frisch, Leontief, and others. But the particular
question that concerns us is as yet so much debated, and in most cases
so imperfectly stated, that we must try to state and answer it1 before we
can discuss in the light of our theoretical expectations the behavior of
such indices as we have, or use them in "correcting" or "deflating"
individual prices or other magnitudes expressed in monetary units.
B. The Theory of the Price Level. — Imagine, for simplicity's sake,
an isolated society without money, the economic life of which merely
consists in the current production of consumer's goods from original
means of production, say, services of labor and of land. These consumers'
goods are sold to the very laborers and landlords who furnish the produc-
tive services. Money and credit being absent2 and unknown, equilibrium
ratios of exchange will establish themselves between all pairs of economic
goods. But there will be no absolute prices. Now express all these
ratios in a common unit, for example, by putting the exchange value of an
arbitrary quantity of any arbitrarily chosen commodity equal to unity.
Thereby the ratios are turned into absolute quantities, which we call
prices. If we want to change from this standard to another, we simply
divide these prices by the price of some quantity of the new standard
commodity in the old system. We do not need, however, any commodity
standard of this sort, but can derive a unit by putting any combina-
tion of equilibrium market values equal to an arbitrary figure, for example,
by ruling that the sum total of market value (prices times quantities sold
in a period of account) of all consumers' goods be equal to 100 units or
100 billions of them. This will uniquely define the meaning of this —
otherwise meaningless — unit in terms of every commodity, just as well as
the choice of a standard commodity would. The obvious practical
1 The solution to be presented derives from Walras. It has already been given by
Professor Francois Divisia (L'Indice Monetaire, Revue d'£conomie Politique, 1925, and
Economique Rationelle, 1928, pp. 252-280), to whom belongs entirely whatever merit there
is in the idea. Reference should also be made to the recent discussion in the Review of
Economic Studies, vol. III.
2 Unless, indeed, money is defined by the criterion of indirect exchange, in which case
many commodities would function as money.
THE PRICE LEVEL 453
difficulties are not relevant to this argument. But, however we proceed
in order to acquire the immense advantage of a unit of calculation and
clearing (a "unit of account")* we must always introduce something that
is arbitrary, both in the sense that the system of quantities of commodities
and exchange ratios does not itself determine it, and in the sense that the
particular decision is — as long as the same equilibrium persists — entirely
immaterial. Also, any change would be immaterial if all prices and all the
other monetary magnitudes could be instantaneously and perfectly
adapted to it.
The social decision which, in order to bring prices into existence, it is
necessary to add to the other conditions that determine the system of
economic quantities need not, of course, consist in any conscious act,
which it would be practically impossible to perform. It may, and
historically did, come about by way of the growth of a social habit, which
evolved the special — and logically rather abnormal — case of the com-
modity standard. What matters here is merely the fact that such an
arbitrary choice, however it comes about and whatever particular form it
takes, supplies the additional equation that we need in order to have
uniquely determined absolute prices. Without a unit's being given in
which to express prices or, as we may now say, without the choice of a
level of prices, these would be indeterminate because it is only their rela-
tions to each other which are determined by the system itself.
Hence the price level, or monetary parameter, is not a mere statistical
aggregate or a mean like the average height of recruits of a given age in a
given population, but a real thing existing independently of the statistician
and to be distinguished from the relations of prices to each other, which
we shall designate as the price system.1 It is not of the same nature as a
group index, and is more than merely the most comprehensive of them.
Considerations of convenience of calculation enter no doubt into the
practical method of its measurement in a secondary way, but not into its
concept. There is, in the same sense, no question of weighting or averag-
ing. Considerations of plausibility have no place at all. The various
tests of index numbers have to be replaced, for our purposes, by the single
question whether, and how accurately, a given formula expresses the
changes in that parameter. It follows, moreover, that the price level,
so defined, is not itself a price and cannot usefully be described in terms of
supply and demand, for these categories apply only within the world of
individual commodities. And it also follows that considerations of
utility or welfare or of the so-called subjective value of money, which may
move in opposite directions for people with different budget combinations,
1 This term is used in the same sense, we believe, by Professor Mills in Behavior of
Prices.
454 BUSINESS CYCLES
are not relevant either to our concept of price level or to the method of
measuring it.1
That social decision then fixes the price level for the equilibrium
state of the economic system which obtains at the moment; but it would
stay as thus fixed only in a perfectly stationary society. In that case,
however, there would be as little meaning to the question what the value
of the price level is as there would be to the analogous question in the case
of a potential. In reality the price level changes all the tim« without any
change in the social decision itself. For practically every change that
occurs in the economic process affects it, and it is extremely unlikely that
changes will occur precisely in such a way as to compensate their effects on
it. In particular, if nothing at all happens in the sphere of money and
credit, the price level will nevertheless undergo variations. And there is
not only meaning but obvious importance to the problem of measuring
them. This problem would be easy to solve, if the price system did not
change, i.e., if prices never changed except proportionally. We could
then read off the change in the price level from the change in any one
price. Unfortunately, however, the price system also could remain
constant only in a stationary society. As a matter of fact, it changes in
time just as frequently as the price level. Hence, the price of any indi-
vidual commodity, as we observe it at any point of time, must be inter-
preted as the result of two distinct components: the price level and the
price system. It is also easy to see that changes in the level can in prac-
tice hardly ever come about except by way of changes in the system —
even as changes in the system in practice hardly ever come about without
enforcing a change in the price level. Yet the two components of
change, however inextricably mixed, are logically distinct. We have not
understood a given change in prices in general or in any single price as
long as we have not quantitatively separated them. Hence the question:
Seeing that we have only the actual prices, is there any means of doing so ?
If the system changed as well as the level, while the quantities of all
commodities remained constant, the variations of total actual expenditure
would exactly measure the changes in the price level, whatever happened
to the individual prices. If quantities also change, the problem becomes,
strictly speaking, insoluble. But here the differential method of analysis
comes to our rescue. Given the usual conditions of differentiability in the
1 This is important to bear in mind, because recent discussion has precisely turned on
the utility aspect in order to define equivalence of different collections of income goods in
different points of time. This equivalence, in turn, is made to yield a criterion for the
change that may have occurred in the "purchasing power" of the monetary unit. From
this standpoint, Professor Haberler has very naturally been led to deny the existence of any
general price level. But all this, important as it is for other purposes, does not concern us
here, and argument on these lines is as irrelevant for our purpose as our argument is for
welfare considerations.
THE PRICE LEVEL 455
(smoothed) time sequence of the relevant magnitudes, we may still
disentangle the rate of change of the price level at any given moment of
time. For this purpose, we start from the expenditure on every one
of our, say, n commodities, equal to its price p* times its quantity q*
(i = 1 • • • ri) bought at any time, form the total differential (rate of
change of expenditure)
and sum over all commodities, so as to get, if total expenditure be E,
{ — n i**n
dE = q4pi + 2 P^dq^
Now, the dq's being the increments, positive or negative, of the quantities
of commodities, ^Lp^dq^ that is, these increments times the "old" prices,
gives approximately that part of the variation in expenditure which is
balanced as to its effects on the level by the quantity changes, and by
which expenditure would have had to change in order to keep prices
constant, if their system had not changed, or at least the price level, if the
system has changed. To put it in still another way, if expenditure had
changed exactly by ^ptdqi9 so that 2qtdpi = 0, that part of the new total
of expenditure would have remained constant which does not buy the new
positive or negative increments and may be looked upon as expended on
the same kinds and quantities of commodities as before. But, with
unchanging kinds and quantities of commodities, constant expenditure
defines identical price levels. And as there cannot be, at one time, two
price levels with respect to the same commodities, the price level would be
the same as before. Hence, for a change in the price level to occur, it is
necessary and sufficient that Sq+dpi ^ 0, and the departure from zero of
this quantity, therefore, approximately measures, by its relation to
2pt#», the change that actually occurred.
We see the principle. It consists in reducing the unmanageable,
but general, case — that of simultaneous change in level, system, and
quantities — to the manageable case of unchanging quantities, which is
always implied in the general case and can be extricated from the rest
if changes are small. This means, of course, that the solution is but an
approximation and that any method based upon it breaks down if the
changes in individual prices or quantities cannot be decomposed into small
ones.1 It also means that comparisons of levels are possible only between
1 It breaks down, therefore, in times of sudden and violent changes, such as extreme
inflation. It does not, however, break down because of the introduction of new commodi-
ties, provided they do not suddenly intrude in quantities that are big with reference to all
the others. How far an index of this nature fulfills Professor Fisher's tests cannot be dis-
456 BUSINESS CYCLES
neighboring points of time, and that the states of things obtaining at
finitely distant points of time cannot be compared directly but only by
way of all intermediate points. In practice, of course, smallness of dis-
tance may be interpreted somewhat less strictly. Dividing monetary
magnitudes by an index of this kind means the elimination from them of
the influence of changes in the significance of the monetary unit i.e.,
of a factor, the action of which makes their comparison all but meaningless
and which is itself meaningless insofar as the really relevant elements of
economic life, quantities of commodities and ratios of exchange, are con-
cerned. This is what is here meant by "deflating** sequences of items
expressed in "current dollars.'*1
Next we have to find out which of the better known formulae may
be considered satisfactory, or approximately so, from the standpoint
of this theory. Of course we are not concerned with any of the many
special-purpose indices, which, as the reader will realize by now, have
nothing to do with measuring the price level in our sense, although the
word level invariably occurs in association with them, and although their
construction may be amply justified by the purposes they are to serve.
Happily we make a very comforting discovery. One of the most
common formulae — common, it is true, more in the theory of the subject
than in practical index making — is the one usually referred to as the
Laspeyres formula. It compares two aggregates: the quantities of the
base period times the prices of the current period and the same quantities
times the prices of the base period. If we denote any price or quantity
of the current period by the subscript i, and any price or quantity of the
base period by the subscript o, the formula is 2ptq0/2p0qo- It is obvious,
and has often been pointed out, that this formula imparts an increasing
bias the farther away we draw from the base period. But if we have, for
example, monthly data, which will in most cases represent "small"
intervals, and if we refer each item to its predecessor as its base ("chain-
method," first effectively advocated by A. Marshall), then we may put
each pi — p0 + dp0 and get, dropping subscripts :
2(p + dp)q _ 1 Zqdp
cussed here. Nor is this question relevant for our purpose. But it may be stated that it
fulfills all the criteria that retain meaning within this theory.
1 Objections to the performance of that operation have been raised by many eminent
authors. They are but too well founded if they rest on the quality of available indices.
They are also well founded if they aim at erroneous conclusions arising from an imperfect
understanding of what "deflating" really means. But the operation remains nevertheless,
as unavoidable as it is theoretically correct if properly understood. It has been pointed
out to the writer, and it may well be true, that the use of the term price level for our concept
may prove misleading. Mr. R. Bryce has suggested the term monetary parameter.
THE PRICE LEVEL 457
This is our own formula. We could also have followed the suggestion of
Paasche and compared the aggregate of expenditure at the current period
with the sum which would have been expended at the base period if at the
then prevailing prices (p0) the current quantities (gt) had been bought.
Hence the suggestion of Knut Wicksell (which he, however, based only on
the absence of a criterion of choice between the Laspeyres and the Paasche
formula), namely, to take the geometric mean between the two (Professor
Fisher's Ideal Formula) is, although not quite in accordance with our
theory, also acceptable provided, of course, that changes are small.
Finally, there is the question which prices to include. If the price level
in our sense is a definite thing, the answer must follow from the theory of
that thing. So it does. We ought to include the prices times quantities
of all commodities and services actually and directly bought by house-
holds, and nothing else. A very simple argument will establish this
principle.1
The price level in our sense is a measure of a property of the system of
economic values. This parameter derives from the relation between the
flow of expenditure and the flow of the things bought by it, and thereby
defines, in a particular way and for a particular purpose, the significance
of the unit of accounting and clearing in terms of commodities and services.
Now this flow of expenditure runs, as it were, through several basins
or economic spheres. For our present purpose we may reduce these basins
or spheres to four: the "markets" of finished consumers' goods, of original
means of production (primarily labor), of produced means of production
(primarily raw materials and machinery), and of titles to income (pri-
marily shares, bonds, and realty) . Expenditure flows through all of these
but each element of it confronts, at any one time, not all but only one of
them. And there is no meaning to a combination of items from different
spheres, or the whole of all the items of different spheres or phases, of the
monetary stream. A variation in one direction of a price in the market of
consumers' goods is not compensated by an equivalent variation in the
other direction of a price in the market of producers' goods. There is, to
be sure, plenty of interdependence between the different spheres, and units
of potential expenditure can be shifted from one to the other. But this is
irrelevant for the arithmetic of the thing. The relevant criterion is sub-
stitutability in the technical sense; we must combine, in order to get the
proper price-level figure, the prices and quantities of all goods which
compete for the sum that actually buys in a given sphere and in a suitable
time interval, and nothing else. If we include less, we may get a change
1 A fuller discussion will be found in the writer's book on money. It is only the
principle that interests us at the moment. In practical work, as we cannot hope to include
everything, the problem becomes one of sampling, and considerations of probability enter
at this stage, although they did not enter into the theory.
458 BUSINESS CYCLES
which is due to a shift of values between rival commodities. If we include
more, we vitiate our result by telescoping different stages or phases of the
monetary process into one. This is what sometimes lies at the bottom of
the argument, which has, however, another surface meaning, that we must
not include wages, except, of course, the wages paid for services directly
consumed, because this would involve "double counting." If we include
"everything purchasable," we get a meaningless heap.
But the only market which exhausts, and does not more .than exhaust,
a complete stage or phase of the monetary process and presents the liga-
men between money stream and commodity stream, is the market of
consumers' goods. The things that form the markets of "original" and
of "produced" means of production are substitutable for each other at
many stages of the productive and commercial process, and no combina-
tion of their prices, therefore, ever displays the changes of our parameter
alone. Nor can we take all producers' goods together, because they do
not fulfill the condition of facing one, and only one, stage of the stream
of expenditure, but are obviously arrayed in successive groups.
In the case of the sphere of titles to income, roughly, the stock exchange
and the realty market, there is a special reason for exclusion. They do not
form a phase of the fundamental stream of expenditure at all and, as we
shall see later on, their pricing process is so different from the pricing
processes in the world of commodities and services that, even for the usual
aims of investigations into the "value of money" and from the standpoint
of the usual theory about the meaning of price indices, inclusion with any
significant weight would yield a result incapable of serving any useful
purpose.
C. The Practical Question. — It would not be impossible to build an
index of the price level in our sense (it could obviously be an index only)
from postwar material. Even for the prewar time information is avail-
able (stretching in some cases over centuries1) that would go far towards
1 Stray data about prices we have, of course, for practically all periods of recorded
history. Data about individual prices, previous to 1780, are sometimes available in long,
continuous series; compare, for example, those in Thorold Rogers' or d'Avenel's works and
Sir William Beveridge's annual series of Wheat Prices in Western and Central Europe,
1500-1869 (see Weather and Harvest Cycles in the Economic Journal, December 1921).
See also Professor Usher's article Prices of Wheat and Commodity Price Indices for Eng-
land, 1259-1930, Review of Economic Statistics, August 1931. For this country several
pieces of work have been published, notably, Bezanson, Gray, and Hussey, Prices in
Colonial Pennsylvania, 1935, G. R. Taylor, Wholesale Commodity Prices at Charleston
1732-1791, Journal of Economic and Business History vol. IV; T. S. Berry, Wholesale
Commodity Prices in the Ohio Valley, Review of Economic Statistics for August 1935; H. M.
Stoker, Wholesale Prices at New York City, 1720-1800, Memorandum 142, Agricultural
Experiment Station, Cornell University; and R. Crandall, Wholesale Commodity Prices
in Boston during the Eighteenth Century, Review of Economic Statistics, June 1934. Back
THE PRICE LEVEL 459
giving us an idea, although too crude for microscopic purposes, of the
behavior of the level. This task, however, is quite outside the range of
the possibilities of the individual worker.
Under these circumstances, we are at present forced to use, and to
perform certain operations on, existing price indices, which for a variety
of reasons bear but a distant relation to what we wish to study. If
we want to cover at least the period from the eighties of the eighteenth
century onward, we must sometimes use (although this has been avoided
wherever possible) annual and quarterly indices, which cannot give correct
contours for the shortest of our cycles or the correct boundaries or extrema
of the longer ones, besides violating our condition that the intervals must
be small. Lack of necessary data about quantities is by itself sufficient
to reduce the value of some indices, at some times of all available indices,
to a mere indication of the strongest features. In some cases the data
themselves and the degrees to which the price quotations used are repre-
sentative of the prices that really ruled (but often differed widely) in the
countries and at the times to which they refer, are open to serious doubt.
Changes in qualities very often present insuperable difficulties. In
other cases technique is deficient beyond remedy, from whatever stand-
point it may be looked at.1 But the old argument of practical workers
that indices tend to give roughly the same picture, however well or faultily
constructed, contains after all some little element of truth, which for us,
to 1680 goes W. B. Weedon, Economic and Social History of New England, 1620-1789,
1891. Of the work done for other countries, Professor E. J. Hamilton's on Spanish and on
French prices is the most important. See also A. P. Usher's series (unfortunately not
published in full) of French Wheat Prices, 1350-1788 (The General Course of Wheat Prices
in France, 1350-1788, Review of Economic Statistics for November 1930), and E. B. Schum-
peter, English Prices and Government Finance, Review of Economic Statistics for February
1938. A long list of contributions could be presented. For the interpretation of such
series a detailed history of the currency is, however, indispensable.
1 On questions of technique and on reliability, as analyzed from the standpoint of
statistical criteria, the reader should consult, from the great literature of the subject, at
least the standard works of Professor Irving Fischer, The Making of Index Numbers,
and Professor Wesley C. Mitchell, Index Numbers of Wholesale Prices in the United States
and Foreign Countries, Bulletin of the U. S. Bureau of Labor Statistics 173, and The Making
and Using of Index Numbers, Bulletin of the U. S. Bureau of Labor Statistics 284; also
Professor Warren Person's The Construction of Index Numbers, Professor Mills' remarks
on the subject in The Behavior of Prices (see, in particular, the instructive graphs on p. 237
and the section on reliability) and the excellent chapter on Index Numbers by the late
A. A. Young in Rietz's Handbook of Mathematical Statistics. Any more comprehensive
study would have to start from Edge worth's famous reports, reprinted in his Collected
Papers. For England, the Board of Trade Index; for the United States, the Bureau of
Labor Index must, as far as the writer has been able to judge, be considered as the best in
the usual sense, as well as in the sense that they approximate our price level more than any
others. On the latter, see H. B. Arthur, The Development of Wholesale Price Measure-
ment by the Federal Government, Review of Economic Statistics for August 1935.
460 BUSINESS CYCLES
it is believed, suffices to justify what we are going to do with them,
provided we watch our step in drawing conclusions.
In one respect there seems to be more reason for confidence than
most students of price-index numbers feel, namely, in respect to the use
of indices of prices at wholesale in lieu of the retail prices for which our
own theory really calls. Owing to the inadequacy of the retail price
material before the war, we have no choice but to use wholesale price
indices, although what is available of the former has been studied.1 But
there are disadvantages inherent in the use of retail price indices which no
critical care or perfection of technique that can reasonably be hoped for in
the near future can be expected to overcome. It is certain retail prices
that are particularly "sticky" in the short run. Some are very "tradi-
tional," and it is here more than anywhere else that change in quality
(also quantity) takes the place of change in price. Moreover, it is diffi-
cult to discover and appraise changes in the services rendered by retailers
to customers and the nature of that composite of ponderable and impon-
derable elements which is what the household really buys, or to follow up
the sometimes tortuous ways of supernormal, normal, and "sale" prices.
Not always, but often, retail trade acts as a bottle-neck that prevents the
stream of commodities from flowing along its course. All this, however
essential for other purposes, is distortion of the true contour lines from
our standpoint. Prices of commodities at wholesale, on the other hand,
may overdraw the picture, but they at least draw it. There is some truth
in Mr. Snyder's saying that they are a picture of "speculation" rather
than commercial reality, and they are certainly likely to display peaks and
troughs, which do not mean much by themselves either as to amplitude or
as to exact location. But that particular kind of " speculation " reflects
the opinion of wholesale trade about imminent reality in the short run,
and the real tendency of things in the long run, free from many frictions,
rigidities, and inertiae. Here, business life does for us a piece of work of
abstraction and analysis that could — and for purposes requiring a higher
1 We are rapidly acquiring valuable series, especially of the cost-of -living type (which,
to be sure, is not quite what we want when investigating the changes of the price level in
our sense). Most of these are in common use, but there are many others, in particular in
some countries not included in this study. Far above everything else, both as to abundance
and reliability of material and as to excellence of workmanship, is Professor Gunnar
Myrdal's Cost of Living in Sweden, 1880-1930 (Stockholm Economic Studies, 1933).
See especially the chart on p. 141, which covers the period from 1830 to 1913. The Canad-
ian Index (R. H. Coats) 1900-1915, some of the material in the English Report of an
Enquiry by the Board of Trade into Working-class Rents and Retail Prices • • • in 1912
(Ed. 5,955, 1913), which among other things gives food-price indices 1900 to 1912 for
fourteen countries, and, for this country, Mr. Carl Snyder's indices and their constituents
have also been very helpful.
THE PRICE LEVEL 461
degree of accuracy will have to — be done on retail series only by laborious
and unsafe methods.
Inasmuch as retail prices are prices of finished products, their relations
to the associated wholesale prices will differ according to the conditions
and policies of the industry or even of the individual manufacturing
(sometimes wholesale) concerns. We find extremely variable patterns
very refractory to generalization and ranging from nearly instantaneous
(even rigid — where wholesale price is a fixed percentage of retail price)
covariation between retail and wholesale prices to what in the short run
almost amounts to independence. In the case of foods (if we exclude
highly processed and branded types) we presumably come nearer than in
any other case to what seems to be the (theoretically) normal state of
things — i.e., that the change of retail price should lag behind, and roughly
be equal in amount to, the change (per corresponding unit) in wholesale
price, either of the object itself or some raw material or semifinished
ingredient of it, a rule that, of course, has to be qualified by taking account
of stocks and expectations. Obviously, this means a smaller percentage
change in the retail price.1 Actually, covariation between indices of retail
and wholesale prices is more in evidence than we have a right to expect.
The reader can easily satisfy himself that cyclical variations displayed
respectively by wholesale and retail indices do not differ sufficiently to
cause serious concern about any of the statements to be illustrated or
proved by the former (see, for example, Mr. Carl Snyder's charts for this
country and England on pages 390 and 397, American Economic Review
for September 1934).
D. Analysis of the Behavior of Price-level Series.2 — This analysis
starts by recognizing that they are — to use the terminology introduced in
the first and fifth chapters — synthetic, systematic, primary, conse-
quential, and that they display a result trend. We have first to form
expectations as to their behavior from the "pure" model, as modified by
the qualifications which constitute the further approximations, then to
1 The above agrees with Professor Bowley's results in Wholesale and Retail Prices of
Food, Economic Journal for December 1913, which to the writer seems still to be the leading
contribution to the subject. Also compare Frances Wood, Construction of Index Numbers
to show Changes in the Price of the Principal Articles of Food for the Working Classes,
ibid., to which the writer feels much indebted.
a The first author to discuss price level in the light of scientific principle was probably
W. St. Jevons. But Tooke and Newmarch's history of prices is also a discussion of the
factors relevant to variations of the price level. Although marred in many particulars by
inadequacy of theoretical equipment and a certain "wooliness," that work is still a mine,
not only of material but also of wisdom, and must be ranked very high. Analysis in the
light of industrial fact is also the great merit of W. Layton's Introduction to the Study of
Prices (1st ed., 1912; 2d ed., 1935).
462 BUSINESS CYCLES
look at our material and compare, and finally to see whether discrepancies
can be satisfactorily accounted for by external factors1 or by defects of our
material. Since these factors and defects are obviously important at all
times and dominant at some, and since, moreover, our process is "inter-
nally irregular," it would be quite unreasonable to formulate our task in
any other way. We must from the outset be prepared for considerable
discrepancies, and all we can hope to find is traces of our process. This
is why the writer feels unable to attach much weight to -exact timing,
particularly to lead or lag of variations in price level as compared with
other elements of the cyclical process. They would have to be very con-
sistent or very considerable to be of real significance in the circumstances
under which it is the economist's lot to work.
Expectations from the pure model are so definite as to make it super-
fluous to elaborate them beyond what has been said in Chap. IV. Price
level should rise in prosperity — under the pressure of credit creation,
which, under the conditions embodied in the pure model, would not be
compensated either by an increase in output or by any fall in "velocity*'
— and fall in the downgrade — under the pressure of autodeflation and of
increase in output — more than it had risen in the preceding prosperity. We
also know that introduction of additional facts by means of successive
approximations does, indeed, tone down, but does not reverse, these
expectations. Existence of unemployed resources in the neighborhood
of equilibrium is one of these facts. But the most important difference
made by the second approximation — the substitution for the two-phase
of a four-phase cycle — adds the expectation that the price level will go on
falling in depression and that this fall should be corrected in recovery.
This does not mean, however, that recovery will carry the price level
exactly to the figure at which it stood in the neighborhood of incipient
1 The writer has not been able to go into international relations — which, from the stand-
point of each individual country, constitute external factors — to the extent required. In
the matter of price levels this lacuna is particularly serious. For many indices (the Sauer-
beck and the Soetbeer indices are outstanding examples; only Dr. Necco's Italian index is
still worse) covariation is tautological, because they are dominated by the great articles of
world trade. But those articles exert their influence on any index and while, from one
standpoint, this is as it should be, seeing that price levels are among the most important
conductors of cyclical effects, from another standpoint it reduces their value as cyclical
symptoms still further. It also invests them with a causal significance that they do not
otherwise have. The extent to which levels have moved in step in our three countries is
best realized by inspecting the chart displaying rates of change, although some readers
will find the differences in behavior still more interesting. International comparison
meets, however, with all but insuperable difficulties. Professor Bowley's work in this
field (London and Cambridge Economic Service) should be mentioned. Finally, the writer
has never been able to do much with the concept of a world's price level. Mention is due,
however, to the work of Professor W. Gehlhoff, Die allgemeine Preisbewegung, 1890-1918,
vol. 149, Part 1 of the Schriften des Vereins fUr Sozialpolitik.
THE PRICE LEVEL 468
depression. Even depression may, but recovery always does, continue
the work of recession by increasing output. In longer cycles also growth
asserts itself. Hence the equilibrium level that will be reached by the
detour of depression and revival, will in general be lower than the
level that obtained immediately before the system embarked upon it.
It should also be observed that, depression being essentially erratic,
it is in each case a question of fact how much there is for recovery to
correct.
In the long swing of the Kondratieff, in particular, short-run fluctua-
tions such as are caused by panics and spirals play so small a role that
there is not the same reason to expect a rise in price level — at least in its
absolute values as distinguished from rates of change — during a Kondra-
tieff recovery as there is to expect it for the recovery phases of the shorter
cycles. The third approximation, which introduced the three-cycle
schema, affects expectations because of the phenomena of interference
that henceforth complicate the picture. Any given phase of any given
cycle then comes under the influence of the simultaneous phases of the
others and may be entirely blotted out or even reversed by them. It is
important to bear this in mind. For while some economists, particularly
those who hold monetary theories of the cycle, may think our expecta-
tions too obvious to be worth stating, others deny that they are borne out
by facts. And the instances to which these may point are mostly —
though not all of them — attributable to the neglect of Kondratieff
effects.1
Charts IV, V, VI, and VII are presented in order to show how far facts
conform to those expectations. Chart IV displays the rates of change of
price level in our three countries, similarities between which are as inter-
esting as are the differences. On the other three charts the reader finds
the graphs (on a logarithmic scale) of the price levels — the indices of
wholesale prices — themselves, together with the graphs of indices of out-
put, of what we take to represent the circulating medium, and of certain
rates of interest, which will be discussed in the chapters that are to
follow. In the workshop of the writer a habit has grown up of referring to
these charts as Pulse Charts. The reason for this is obvious. Little
though the writer thinks of the explanatory value of aggregative theory,
and far though he is from claiming barometric value for the four con-
stituents of these charts, they nevertheless give a rough picture of the
economic process in time and, in a sense, sum up what we have to account
for by our analysis. Also the picture is, as far as it goes, complete: none
of the four constituent curves could be left out, but no other is logically
1 The above refers only to denial of the association of rising prices with upgrades and of
falling prices with downgrades. If those authors only mean to discount the causal role
of the price level, we are, of course, largely in sympathy with them.
464
BUSINESS CYCLES
g <^
g 3
S -g
I
1
ooopoooo o l'
•— • CJ K> CM •" T*
4* I 4 I + •*• I
THE PRICE LEVEL
465
necessary in order to convey what meaning there is in the variations of
aggregates at all. If the writer had to construct an index of business
conditions, this is what he would offer.
Naturally, we shall first look for the result trend which our analysis
leads us to expect. For since in the downgrades of all cycles the price
level must, barring interference by opposite phases of other cycles,, fall
more than it rises in their upgrades, capitalist evolution produces a long-
run (or "secular") tendency of prices to fall. This downward result
17851790 1800 1810 1820 1830 1840 1850 1860 1870
CHART V. — British prewar "pulse" (see Appendix,
1890 1900 191015
p. 1052).
trend embodies the method by which the capitalist mechanism diffuses
the fruits of industrial improvement over the masses of the people, and
characterizes the specifically capitalist "road to plenty." It is an
illogical method, no doubt, which records increase in real income in a
way that may be likened to measuring the growth of a child by leaving
the number of inches constant and increasing the size of the inches
instead. And other methods of social accounting could be devised which
might achieve the same results without each time creating the danger of
the system's sliding off into depression. Experience tends to show,
however, that neither capitalism itself nor the social institutions asso-
BUSINESS CYCLES
elated with it, democracy among them, can work efficiently and with
comparative smoothness except on a falling trend in prices.1
The price-level curves in charts V, VI, and VII display, in fact, a
falling descriptive trend.2 But although this descriptive trend is what
1830 35 1840 45 1850 55 I860 65 1870 75 I860 85 1890 95 1900 05 1910 15
CHART VI. — United States prewar "pulse'* (see Appendix, p. 1053).
is left in our figures of the result trend, the former does not of course
render the latter. The two differ by the effects of external factors.
1 That statement rests on the opinion that all alternatives that are politically feasible
carry with them other effects which in one way or another tend to upset the working of the
system. It should again be remembered, however, that what has been said does not apply
to the fall of prices in depression.
2 In retail prices, particularly if we include rent, it is much less pronounced. But no
index, however constructed, has, as far as the writer knows, succeeded in effacing or revers-
ing it. It is of course quite incorrect to include wages.
THE PRICE LEVEL
467
Government deflations have never gone far enough to counter-balance
the corresponding government inflations, which in our countries and
period have been mainly, though not exclusively, those of the Napoleonic
time and of the American Civil War. Demonetization of silver, on the
one hand, and the impact of Australian, Californian, and South African
gold discoveries, on the other, must have exerted some influence. .The
spread of deposit banking acted in the same direction as the gold dis-
coveries. Protectionism and immigration of capital also have, in some
cases, been responsible for rising prices, or have prevented a fall that
would have occurred otherwise. Theoretically the descriptive trend of
the sum total of money incomes should measure the net influence of all
65 1870 75 1880 85 1890 95 1900 05 1910 13
CHART VII. — German prewar "pulse" (see Appendix, p. 1054).
these external factors. Unfortunately, however, the evolutionary
process itself tends to bring about not only the recurrent expansions and
contractions of deposits that are described by our pure model, but also
those lasting expansions which we took into account in our successive
approximations and in consequence of which money incomes will display
a long-time tendency to increase. Hence, these expansions become part
of our process, and the method alluded to, therefore, cannot be expected
to serve as a means of separating the result trend from the effects of other
factors. Moreover, there are other reasons for this. Structural changes in
habits of payment are also induced by our process itself. Increase in the
supply of monetary gold cannot be considered as a wholly independent
factor. Government inflations and 'deflations and the policies of which
they are elements have many effects on the economic organism besides
468
BUSINESS CYCLES
influencing incomes. A very interesting research program suggests
itself here. It is, however, safe to say that, for our countries at least,
the net effect of external factors has been to counteract and not to inten-
sify the influence of the result trend.
It stands to reason that external factors affect not only the result
trend but also the cyclical behavior of price-level series. Moreover, the
facts that our figures do not express at all accurately the level of prices
in our sense and that they even do not measure correctly what they could
be expected to measure are likely to make themselves felt still more in
the case of cycles than they do in the case of the secular tendency.
Finally, there is always so much "slack" and underemployment in the
1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910
CHART VTII. — Nine year moving averages of price indices (see Appendix p. 1054).
system and so much room for the play of a multitude of factors that the
effects on the price level of each cyclical phase may be slow to emerge and
can easily be drowned in the effects of the succeeding one. It should be
equally clear that this does not invalidate our analysis.
In the case of the Kondratieff, some readers might think it a waste of
space to prove the existence of those protracted periods of rising and of
falling prices which stand out clearly enough and are seen at first glance
to correspond roughly with the upgrades and downgrades of that long
wave. In fact, all we need to do in order to prove this is to look at the
pulse charts or to take a nine-year moving average of American, British,
and German indices of prices at wholesale (see Chart VIII). It cannot
be too often repeated, however, that mere shifts in the price system,
THE PRICE LEVEL
469
though, they do not per se influence our price level, do influence the indices
we have. Hence individual prices may, besides their legitimate influence,
acquire an illegitimate one as well.
We see fairly smooth wavelike lines that display two units of about
equal length and the beginning of a third. Chart IX, which gives the
variations of American wholesale prices treated by a simplified freehand
adaptation of Professor Frisch's method, discussed in Chap. V, presents a
similar picture. Recalling the testimony of industrial history, we shall
associate the first unit with the processes usually referred to as the indus-
trial revolution. But the height and location of the maxima, being
A 3 YEAR ON A 2 YEAR MOVING
AVERAGE (GRAPHICALLY
~ DETERMINED)
CURVE THROUGH
INFLECTION
POINTS OF 50 \\
YEAR CYCLE -\\
:ffiVE THROUGH INFLECTION
POINTS OF 22 YEAR CYCLE
1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1919 1920
CHART IX.— (See Appendix p. 1054.)
obviously conditioned by the Napoleonic Wars, prove nothing. The
English maximum (Silberling index) occurs in 1814,1 after which "defla-
tion" due to normalization of government finance, superimposing itself on
1 Mrs. Gilboy (Review of Economic Statistics, 1936) gives 1818, but since she uses harvest
years, there is really no difference. Like Jevons before us, we should have expected the
maximum to occur earlier, the more so because English war finance, with relapses, grew
steadily sounder after the turn of the century, though Mr. Silberling's figures do not quite
agree with this. Dr. E. B. Schumpeter (Review of Economic Statistics, Feb. 1938) in fact
established that prices of domestic commodities fell from 1800, which would agree well
with our cyclical schema. It must be remembered, however, that the influx of precious
metals from Spanish South America, which before the wars of the French Revolution was
over 7 million pounds per year, began to fall later on. It was about 5 millions in 1825 and
only 4 millions in 1829.
470 BUSINESS CYCLES
the autodeflation we should theoretically expect, possibly blurs the picture
for about half a dozen years. We cannot eliminate this disturbance by
using gold instead of paper prices, for gold prices also are influenced by
inflation and deflation and by the events of which inflation and deflation
are but the monetary garb. The failure of our figures for price level to
rise in the eighties of the eighteenth century may, in the case of England,
be due to imperfections of our material. New evidence of increase in
prices of many important commodities points in that direction. In
America, the beginning of the Kondratieff is, as we have seen, doubtful,
owing to the Revolution and its aftermath.1
The third and smaller hump that we observe in the chart of nine-
years moving averages presents more delicate problems. The occurrence
of such a fluctuation is not astonishing in itself and proves nothing against
the three-cycle schema, because it may be but an effect of superposition
(see Chart I and the top curve in Professor Fisher's chart in Business
Cycles as Facts or Tendencies, p. 6, in the volume in honor of Professor C.
A. Verrijn Stuart). But the amplitude is unexpectedly great for a Juglar
belonging to the recovery phase of a Kondratieff. Moreover, price levels
continued to fall far into the forties, according to some indices much beyond
our date for the rise of the prosperity of the second Kondratieff. We
have met the same difficulty in our historical discussion.2 Of course,
it may mean no more than that our schema is faultily designed. This
may well be so. But, as stated in Chap. VI, the writer feels inclined
to believe that "reckless" banking practice, which undoubtedly was an
outstanding feature of the thirties in both America and England, inten-
sified the prosperity phase of the last Juglar of that Kondratieff and
induced a wave of speculation of unusual violence which, in turn, accounts
for the depression of unusual severity, and for underemployment at the
beginning of the second Kondratieff. This explanation seems partic-
ularly plausible in its application to the price level, while political troubles
1 Mr. H. M. Stoker speaks of a postwar depression that lasted for eight years (Memoran-
dum 142, Cornell Agricultural Experiment Station, 1932, p. £04). This hardly accords with
historical evidence (see Professor W. B. Smith's comments in Smith and Cole, Fluctuations
in American Business, 1790-1860, 1935, p. 12). But it is true that existing figures do not
indicate a rise in price level before 1792. Miss Crandall's "special" index of Boston prices
does rise from 1788, but it consists only of molasses, rum, and fish. The situation was,
as we have seen, prosperous in 1790, and became violently booming immediately afterward
(see Chap. VI, Sec. D). The Hamilton policy, though favorable to prosperity, was not
favorable to an increase in prices. Also, the virgin environment offered plenty of under-
employed resources.
2 Therefore, Kondratieff himself dated the second long wave from 1849. Our view is,
however, supported by the authority of Spiethoff, who has an Aufschwungsspanne 1842-1873
on the criterion, perfectly valid from our standpoint, that years of prosperity (he counts 21)
predominate over years of depression (10).
THE PRICE LEVEL 471
would, as has been pointed out by Sauerbeck himself (Journal of the Royal
Statistical Society, 1886, p. 648 l), no less plausibly account for the further
fall, 1848 to 1851, which is, however, perfectly regular within the course
of the Juglar. Moreover, in England, where the fall of the price level is
most obvious, introduction of free trade must have had some effect on
prices, and Peel's Act may have exerted some restrictive effects on credit
creation, as in fact it was intended to have. If there is anything in this,
we must conclude that we have simply a case of a primary cyclical element
failing to behave according to expectation because of a combination of
counteracting circumstances. That this did not prevent a prosperity
phase in our sense from starting and running its course is clear in any
case : the English railway mania testifies to that with no uncertain voice.
The behavior of certain indices seems to lend some support to this
interpretation. In Germany, where the hump of the thirties, though not
absent, is much less in evidence than in England and America, prices
began to rise in 1845 (see the Berlin Institut's new index). In France
there was a sharp upturn of the price level, at least from the beginning of
1844, possibly from the end of 1843. For the United States, Professor
Cole's 38-commodity index (Review of Economic Statistics for April 1926;
1834-1842 = 100) gives a maximum for 1836 (130) and a minimum for
1842 (72), after which a rise set in that lasted up to the point where we
locate the turn of that Juglar (1847). Professors Warren and Pearson's
all-commodity index reaches its trough in 1843 and then begins to rise. So
do their indices of metal and metal products and of building material.
It does not seem unreasonable to describe this as a tendency toward
increase, dwarfed by the circumstances referred to above.
Recalling what has been said on the subject in Chap. VII, we find
that in all three countries (also in France, according to the index of the
Statistique Generale de France) a peak occurs in the middle of the fifties
that may be identified as a perfectly regular Kondratieff turning point
Dates of Minimum for Sauerbeck Group Indexes
All commodities . . . 1849, 1851 (equal)
Total materials 1848, 1849 (equal)
Food 1851
Corn, etc.. . 1851
Animal foods 1850
Sugar, coffee, tea 1848
Minerals 1851
Textiles 1848
Sundry materials 1849
See Journal of the Royal Statistical Society, 1886, p. 648, Ten-year moving average touches
its minimum in 1848. However, the Sauerbeck index is no ideal guide.
47£ BUSINESS CYCLES
But the fall, which we should then expect to continue, with interruptions
due to the prosperities of the shorter cycles, for at least the whole of the
Kondratieff recession and depression phases, was everywhere, except in
France, checked about 1858. To about 1863 the Juglar recovery and pros-
perity account for that, when the German price level in fact begins to fall.
That turning point occurs in England one year later, and in the United
States (if we take gold prices) two years later. This and the violence —
not the fact — of the rise in prices, which in England" and Germany
occurred from 1870 to 1873 — both gold and currency prices moved up but
moderately in this country in 1871 and the first half of 187& — afford the
only occasion for invoking the influence of gold production in explanation
of movements contrary to expectation, although on others that influence
may have intensified tendencies independently induced by different
factors. We shall understand, however, that economists who put their
trust in the formal properties of a graph will look upon the whole
segment from about 1849 to 1873 as a long-time upgrade. But since we
can easily account for that peak — which, to repeat, is irregular in nothing
but height — and since the reaction was so prompt and so severe, the
thesis that the underlying tendency was downward from the middle
fifties is perhaps not indefensible. If so, the consequence that follows
for the value of formal methods of analysis is truly discouraging. Price
levels then continued to fall, not only through the depression but also
through the recovery of the Kondratieff, though for all our countries
it is possible to speak of a fall at a decreasing rate, which in the case of
Germany almost vanishes by 1886. Thus Kondratieff recovery again
failed to bring about a recovery in prices. The reasons have been men-
tioned at the beginning of this section. The rise of the third Kondratieff
is clearly marked and particularly normal in this country (1897).
We may sum up by saying that the great waves of economic change
recorded by history show in the behavior of the price level, but that the
association is so imperfect as to make it highly unreliable for purposes
of diagnosis or prognosis. Since existence and adequacy of the disturb-
ances that we hold responsible for that imperfection can in each case
be established from independent historical evidence, the fact should not
be recorded against our model. Among them, monetary disorders, which
in particular account for the outstanding peaks, are by far the most impor-
tant. But again the question whether we have done justice to the gold
factor crosses our way. The problem of its possible role in the causation
of business cycles having been dealt with in Chap. IV, Sec. E and Chap. VI,
Sec. C, it remains to touch upon the problem of its causal importance for
the price level.
Professor Gustaf Cassel, whose advocacy of the time-honored view
that the price level is dominated by gold production in relation to output
THE PRICE LEVEL 473
or, more generally, economic activity1 has been instrumental in starting
the discussion anew, compared world stock of gold for 1850 and 1910, in
which years the Sauerbeck index stood at approximately the same figure,
concluded that the average increase in gold stocks between those years
(2.8 per cent) was "normal" in the sense that it was sufficient to keep the
price level constant, and compared variations of the ratio of the actual to
the normal quantity of gold with the variations in price level that had
occurred, extrapolating to 1800. Mr. Joseph Kitchins corrected figures,
took account of Indian, Chinese, and Egyptian hoarding, restricted gold
stocks to monetary gold stocks, and arrived at 3.1 per cent as the normal
yearly increment. Fit was greatly improved and the obviously desirable
precedence of variations in that ratio, as against variations in price level,
was not wanting. Even so, the method not unnaturally met with adverse
criticism.2 Besides, it might be urged that the growth of deposit banking
must have alleviated the "scarcity" of gold after 1873 and extension of the
area of the gold standard, the plethora of gold after 1896. Much further
research will have to be done before we can speak with confidence on these
matters; but for our purpose we need not go into them. We will grant
for argument's sake more than we believe should be granted and accept
as it stands the evidence offered for the effects of gold on price level. It
does not follow that the Kondratieff wave in price level is simply due to the
variations in gold production. On the contrary, it is clear — since accord-
ing to that theory price level is the result of variation of monetary gold
stocks (which, let us note in passing, are still more a function of business
situations than total gold stocks) and output of commodities, and since
variation in the latter result, in turn, from the working of our process —
that whatever the behavior of gold, unless it should happen to be exactly
compensatory, the finger prints of the Kondratieff must show on price-
level graphs, although more or less blurred by gold production.
1 Having already presented his method before the war, he substantially repeated his
argument in the League of Nations Rapport Provisoire de la Delegation de I' Or du C omit 6
Financier, annexe X. Mr. J. Kitchin's work on the subject was also done for the same
delegation.
2 The reader finds an exposition and defence of the method in W. Woytinsky, Inter-
nationale Hebung der Preise als Ausweg aus der Krise, 1931, and a good presentation of the
case against it in Mr. J. T. Phinney's article Gold Production and the Price Level. The
Cassel 3 % Estimate, Quarterly Journal of Economics, 1933. See also Mr. Tucker's paper on
Gold and Prices, Review of Economic Statistics, 1934. Fit is even better than the sponsors
of the method claim, for they fail to take account of the important role silver played in the
first half of the century where, over the interpolated range, the fit is less satisfactory.
Overlooking this, Mr. Woytinsky tries to improve the fit by assuming a smaller rate of
progress during that time, in order to pull down the curve of the ratio between actual and
normal gold supply until it goes through the curve of the price level. But there is as little
necessity for this "correction** as there is warrant for it.
474
BUSINESS CYCLES
In spite of all the objections that may justly be raised against the use
of moving averages, we will take the deviations from our nine-year moving
averages in order to get some idea of the Juglars. Chart X exhibits the
result. Although with characteristic differences in amplitude, they seem
to the writer to stand out rather well for all three countries. There are,
in England and the United States (but we may with practical certainty
assert the same thing for Germany), six of them to the only complete
Kondratieff covered by the chart. The reader should -mark them off
and form an opinion about the question how far there would be any sense
Per Cent
+20
+10
0
-10
+30
+20
+ 10
0
-10
+20
+ 10
0
-10
-20
-30
BRITAIN
^
A,
\r
V
/Wr
w/
V
UNITED STATES
V
V
1780 1790 1800 1810 1820 1830 1840 I860 I860 1870 1880 1890 1900 1910
CHABT X. — Percentage deviations of price indices from their nine year moving averages
(see Appendix p. 1055).
in trying to measure average amplitudes — bearing in mind the Kondra-
tieff phases from which they rise and to which they have, in each case,
to fall — and to base conclusions on such averages. In passing, we will
advert to Mr. Philip Wright's experiment (Moore's Economic Cycles,
Quarterly Journal of Economics, 1914-1915, p. 635), which consisted in
correlating successive items. Correlation was found to be highest for
items distant from each other by 9 or 10 years. The coefficients are low —
.40 for 9 years and .35 for 10 years — and the procedure is otherwise open
to criticism. But the result accords well with our findings. Also, it
should be observed that some of the objections to applying correlation
analysis to time series do not hold in this case. Professor E. B. Wilson's
periodogram of prices shows only an unsatisfactory hump for a 109-month
THE PRICE LEVEL, 475
period, but it may be significant that other humps also occur at periods
of about twice and three times that length.1
Charts X and IV may be consulted in order to get an impression of the
Kitchins. No doubt is possible as to the presence of shorter fluctuations
within the Juglars. It seems fairly clear also that there are about three
of them to each Juglar. They are highly irregular in amplitude, some-
times showing by a mere kink in the curve. This can however always be
explained by the influence of underlying cycles, which — as mentioned
above — may go so far as to iron them out. Their period is less irregular
and does not stray far from the average.
E. Group Prices. — An index of a family of prices differs, toto caelo,
from what we have been discussing in the preceding sections. There is
no objection of course to calling it a level, but it is not a parameter char-
acteristic of the system as such or of the monetary ligamen. On the one
hand it is no "real thing" existing irrespective of the doings of statis-
ticians : it owes its existence, not only its presentation, to them. On the
other hand, it remains a price, subject (if the group is not too big2) to
the ordinary logic of prices, and should simply be called a composite price
or a composite price relative, as the case may be. In the most favorable
instances the matter involves no other problems than are always implied
in speaking of "the" price of a not perfectly homogeneous commodity or,
say, the composite price of finished steel. In other instances the object
is to combine commodities the prices of which display some feature com-
mon to all of them, without attempting to find anything that will bear
interpretation as a composite price; we shall look upon indices of group
prices in this light.
One type of case may be instanced by the familiar indices of prices of
textiles, metals and metal products or, if the relationship is on the side
of demand, by indices of prices of building materials. Another type
consists of groups such as the group of Sensitive Prices,3 so well known
to students of forecasting in this country and in Germany. This is
merely an assemblage of prices empirically found to fluctuate more
1 The periods of the Juglars do not seem to be affected by the Kondratieff phases; but
duration of the Juglar phases may be, i.e., prosperities may be not only more marked but
also longer in Kondratieff prosperities, than in Kondratieff downgrades. This is what has
been observed by Professor Spiethoff. The writer prefers to say that in the former cases
recessions look more like prosperities and in the latter, more like depressions, and not to
stress difference in duration. But this is little more than a different way of expressing the
same thing.
2 If it is too big, the Marshallian supply-and-demand apparatus loses its meaning.
We may still speak of a composite price, but only the roughest of popular meanings attaches
then to the concepts of supply and demand, and they will no longer bear modern refinements.
3 See, for instance, Professor W. M. Persons' index in Forecasting Business Cycles,
p. 93, et seq., also the index of the Institut filr Konjunkturforschung.
476 BUSINESS CYCLES
violently than others. It serves a useful purpose, but since for us the
wholesale price index is quite sensitive enough, we may dismiss this group
with the remark that its divergence from the general index gives what so
far has been the relatively most successful measure of cyclical variation
of price dispersion. Price indices of durable and transient, of domestic
and foreign, of monopolized and not monopolized, of raw and processed,
of basic and other commodities, of foods and nonf oods, are instances of the
application of what may fairly be called a distinct method of analysis.
Regional group prices are really levels. Then there is the great division
between producers' and consumers' goods prices, which comes near to
defining entire spheres of the monetary process, provided we look on both
the producers' and the consumers' goods prices included as random
samples.
Finally, we might combine, with suitable weights, an index of prices
of raw materials with, first, a composite of an index of prices of equipment
and of building costs, and second, an index of wage rates corrected for
changes in product per man-hour. Thus we could get an index of costs
and compare it with an index of prices of finished products. But unless
considerable means were invested in this project, results would be exposed
to such serious errors that we could not trust them beyond what we may
in any case infer from indices of prices of raw materials, wages, and so on,
taken separately.1 More progress has been made, since the war at least,
in the construction of indices of the prices of wage goods — the cost-of-
living indices.
Suitably chosen group prices might be expected to show, in the course
of the cycle, characteristic variations relatively to each other. So they
1 The problem is different, of course, according to whether it is desired to build a " social"
index of cost of production (or cost of doing business in general) or whether the goal is the
more modest one of deriving an index of costs for a given industry or, still more modestly,
a given concern. Professor Mills' investigations on cost in his Economic Tendencies in
the United States, the cost indices of the late G. T. Jones (Increasing Returns, posthum-
ously published, 1983), and also Professor Mitchell's pioneer work in his great book of 1918
must be mentioned. The theoretical bases for further advance have, it seems to the writer,
been laid in Professor Leontief's Studies on the Elasticity of Supply (Weltwirtschaftliches
Archiv for January 1982), which paper also presents interesting applications to the iron and
steel industry of the United States. An example of an index of cost for one concern is given
in K. Ehrke: Uebererzeugung in der Zementindustrie, 1858-1918 (1983; the method is
Dr. Schneider's). Cost investigations, of course, abound since the war; but they rarely
cover more than a few years. Construction costs are a particularly bright spot. See,
for example, the index of the Federal Reserve Bank of New York and that of the American
Appraisal Company. Compare L. J. Chawner, Construction Cost Indexes as Influenced by
Technological Change and other Factors, Journal of the American Statistical Association
for 1935, particularly, Chart V on p. 57, and W. D. Conklin, Building Costs in the Business
Cycle. These studies are mentioned here, although they deal almost exclusively with war
and postwar material, because some of the questions of principle involved can hardly be
dealt with on prewar material.
THE PRICE LEVEL 477
do, of course. These variations can be brought out by dividing all of them
by the index of the general price level1 and can be used for the purpose of
measuring certain types of cyclical price dispersions and disequilibria.
They are of considerable importance for the higher approximations of a
more refined analysis and for any quantitatively exact picture of cyclical
situations. But if we look at Chart XI, which is sufficient to illustrate
the one point we wish to make and which the reader can easily supplement
by other such graphs familiar to everyone, it is not that aspect which
strikes us first. What stands out is, on the contrary, the covariation:
compared with it, all there is of difference in form, amplitude, and period,
and of lag is distinctly secondary.2 Let us pause for a moment to con-
sider that broad truth which, though only broadly, yet asserts itself
consistently.
From our analysis of the cyclical process of evolution it follows, indeed,
that cycles are not satisfactorily described as aggregative movements
that leave structural relations within the system untouched. It is of
the very essence of that process that it remodels the structure of the
system. But it does not follow that comparison of the cyclical behavior
of group prices must show this. We will divide all the groupings of
prices that have been mentioned above into three classes. To the first
we assign those group prices the constituents of which are related by
virtue of affinity between the commodities — like group prices of textiles,
electrical apparatus, and so on. To the second belong group prices
formed from prices of commodities that have in common some character-
istic of organization or marketing; prices in large-scale industries or
monopolized industries afford examples. Into the third we put group
prices built from constituents which dwell in distinct "stages" of the
economic process — producers' and consumers' goods being a typical
example.
Now, we shall expect that, both within the shorter cycles and in the
long run, group prices of the first class will behave differently according
1 This has been done very often, particularly in order to express variations in the "pur-
chasing power" of the products of agriculture (the German word Agrarschere has no good
English equivalent), see, for example, Kondratieff, Die Preisdynamik der industriellen
und der landwirtschaftlichen Waren, Archiv fur Sozialwissenschaft, vol. 60. Mr. Carl
Snyder made extensive use of that device in his papers on the Structure and Inertia of
Prices (American Economic Review for June 1934) and on Commodity Prices versus the
General Price Level (ibid., September 1984). In the facts and in some results the writer
entirely agrees with him. The above may even be considered as an interpretation of some
of his findings. However, the obvious dangers of the method, particularly when applied to
groups, should not be overlooked.
8 This would be different if, following some students, we took the price of shares (as
represented, say, by the A-CUTVG of the Harvard Barometer) to mean price of real capital.
But to the writer this seems entirely inadmissible.
478
BUSINESS CYCLES
to whether or not they are dominated by prices of industries that happen,
in each interval of time, to be innovating. There are indeed many
qualifications to this. New commodities lack, at their start, a standard
of comparison by which to characterize their relative behavior. As far
1840 45 1850 55 1860 65 1870 75 1880 85 1890 95 1900 05 1910 13
CHART XI. — United States group prices (see Appendix p. 1055).
as innovations occur in highly finished goods, the price variations easily
escape record, both for statistical reasons and because such innovations
often result in the offering of better qualities for the same price. Even
where innovation does influence quoted price, the effect may to a large
THE PRICE LEVEL 479
extent be lost by grouping. Moreover, innovation in most cases also
acts on other prices as well as on the prices of the innovating industry,
and entrepreneurial activity shifts from one industry to another, so
that we must look for its1 influence — particularly its long-run influence
— to the price level rather than to group prices, which for those reasons
will tend to "catch up" with each other. But we may still expect
innovation to show and so it does. If the reader inspect Mr. Snyder's
charts (in the first of the two papers quoted in our footnote, pp. 189 and
190) he will not find it difficult to associate the behavior of the group
prices of chemicals and metals or (in the late seventies and eighties) of
fuel and lighting with innovation, or the behavior of the group prices of
building materials and of hides and leather with (comparatively speaking)
the absence of it. Much more striking instances (textiles at the beginning
of the nineteenth century, power at the beginning of the twentieth, for
example) could easily be given. The group price for farm products, far
from contradicting the principle, really strengthens the evidence for it.
There was plenty of innovation, both in the production and in the trans-
portation of farm products, but since both opened European markets
for American exports, it is in perfect accordance with our point of view
as well as with general theory that this group price did not fall much in the
United States. In England it fell. In Germany it did not, but this of
course is due to the protection to agriculture which set in at the end of the
seventies.
The cyclical behavior of group prices of the second class is conditioned
by the cyclical effect of the properties that the constituents of each group
have in common. Monopolistic pricing is an example for such properties
and illustrates the proposition that they primarily influence the response
to the impact of the active elements of the cyclical process (see Chap. X).
But in group prices of the third class we shall expect fundamental
covariation. Although entrepreneurial activity impinges on particular
spots in the system, the effects of the increase in producers* expenditure
in prosperity and of the decrease of it, in amount or rate, toward the end
of prosperity and in recession, will make themselves felt very quickly
all over the system. Increasing wages, in particular, are promptly spent
in prosperity (qualifications of secondary, if yet considerable, importance
will be noticed at a later stage) and under the influence of optimistic
expectation (greased in many cases by ready availability of consumers'
credit) expenditure may even outrun actual receipts, as pessimistic
expectation may cause consumers' expenditure to decrease by more than
1 This should be emphasized. Many economists are in the habit of arguing as if innova-
tion affected relative prices only (as, in fact, it does directly) and as if, hence, movements of
the price level could have nothing to do with it. But it is easy to see that these movements
cannot be treated as an independent component of individual prices.
480 BUSINESS CYCLES
the actual decrease in receipts, as soon as a depression gets under way.
There is, hence, no theoretical reason to think that consumers' goods
prices must always or necessarily or significantly lag behind producers'
goods' prices. Nor is there any reason why they should precede. From
our theory (Chap. IV Sec. A) we should rather expect almost synchronous
movements in the same direction in all sectors of the economic organism,
and differences, in a big country, to be more regional than "stagewise."
That this is so, is the most important result to be gleaned from any study
of group prices of this class. The facts do not lend much support to
those views about the cyclical mechanism from which characteristic
sequences would follow, and least of all to the theories about savings
deflecting "purchasing power" from consumers' goods to producers'
goods.1
This does not mean that deviations from synchronous parallelism are
either unimportant or uninteresting. They are more important in the
first two classes of group prices, but they are not absent in the third class.
It would be most astonishing if there were no such deviations, because
each group price combines such different patterns of action and response
that it would be little short of miraculous if the outcome were equal and
synchronous percentage change for all. But the point is that we cannot
generalize about those differences, which have their roots in the peculiar-
ities of the individual commodities entering each group and in the con-
ditions of each cycle and, therefore, teach little about the fundamental
mechanism of cycles in general. Thus the reader will see from our chart
that the producers' goods price composite, although it displays sub-
stantially the same rhythm as the consumers' goods price composite,
shows greater amplitudes and generally also some precession.2 But
1 It is perhaps worth mentioning that one of the many mistaken interpretations of the
analytic schema presented in this book was to the effect that, according to it, prices of con-
sumers' goods should be the first to fall and the ones to create "the trouble." This is a
complete misunderstanding of the proposition that in the end the cyclical process of evolu-
tion means increased real income to all classes, conveyed to them primarily through falling
price levels.
2 One of the first attempts at systematic comparison between producers* and consumers'
goods prices was that of Professor Mitchell (see Bureau of Labor Statistics Bulletin 178).
The period between 1890 and 1918 affords instances for the above: the wider fluctuations in
producers' goods prices at wholesale stand out clearly enough. They fell sharply to 1894
and less than consumers' goods prices to 1895. Then they rose in 1896, while consumers'
goods prices were still falling; after a setback they rose together with consumers' goods
prices, overtaking them in 1898. They did not fall in 1908 and 1904 when consumers'
goods prices did. In 1909 consumers' goods prices rose, while producers' goods prices
were still falling. In 1913 the situation was similar. But even as far as producers' goods
prices can be said to lead, it will be more true to reality to stress the fact that the people
who buy them have, as a class, a wider range of foresight, rather than the fact of their
position in the productive process.
THE PRICE LEVEL 481
this is not wholly due to greater distance from the demand of ultimate
consumers, which, though it actually is comparatively steady, yet
fluctuates quite appreciably in the course of cyclical phases. There are
raw materials that have only a short way to go before reaching households
and yet fluctuate as much as some others that are a long way off. Com-
parative steadiness of consumers' goods prices, such as it is, is to a large
extent a result of a quite different factor, viz., that consumers' goods
prices, even if not retail prices, include more branded and serviced
articles than the producers' goods group. Hence, we must not trust too
much to mere distance from consumption, but also look for other explana-
tions, such as durability, which accounts for the violent fluctuations in
the group price for metals and metal products. Certainly entrepreneurs'
demand impinges conspicuously there. But it also impinges on labor
and thus on consumers' goods, and there is no cogent reason for assum-
ing, a priori, that the one effect must be stronger than the other.
From the standpoint of our model of the cyclical process this is largely
accidental.1
While these considerations tell, as in fact they are intended to, against
theories which assume systematic relations between group prices that do
not exist at all or do not assert themselves as clearly as they would have
to in order to prove anything for these theories, they do not tell, nor are
they intended to tell, against a proposition dear to some "endogenous"
or "self -generating" theories, viz., that as prosperity wears on, costs of
doing business increasingly encroach upon profits. We do not consider
this, taken by itself, to be a satisfactory explanation of the turn of
prosperity into recession. Nor are we able to accept as relevant all the
factors that have been listed as contributory — losses due to bad debts,
for example, hardly increase before, or independently of, the turn. But
we do not deny the reality of the mechanism, which indeed forms a part
of our model. Many "old" firms, as pointed out in Chap. IV, will be
inconvenienced from the outset by the rising costs of labor, materials
(which however again include labor), credit, and so on, and most "old"
firms — all in fact that do not receive more than their share of the expendi-
ture induced by entrepreneurial activity — will get into difficulties as soon
as entrepreneurial activity begins to slacken, even apart from the effects
of direct competition by "new" plants and from the difficulty, frequently
* * Covariation, tempered by peculiarities of the industries involved, is also the outstand-
ing feature of the movements of foods and materials. Here, if anywhere, should we expect
neighborhood of the sphere of consumption to assert itself, since the demand for food is
certainly much more cyclically stable than the demand for anything else. It does assert
itself, but hardly ever to the point of discrepancy in direction, the cases of which are, more-
over, all explainable by circumstances that have nothing to do with the cyclical process.
See, for example, the Sauerbeck Food and Material indices, Journal of the Royal Statistical
Society, 1910, p. 316; 1929, p. 239.
482 BUSINESS CYCLES
present in the case of finished commodities, of increasing a price which is
fixed by custom, business policy, or public authority. As Professor
Mitchell has shown, this will, for many industries, result in buying prices
of materials gaining on selling prices of products. This, however, has
only a distant connection with the relative movement of producers'
and consumers' goods prices.
CHAPTER IX
Physical Quantities. Employment
A. Individual and Composite Quantities. — As in the case of prices,
we have data about quantities produced, forwarded to mills, transported,
exported, imported, bonded, subjected to payments of excise, sold to
consumers (sometimes put equal to quantities produced plus imports
minus exports), visibly available or reflected by such indicators as
spindles active or active spindle hours, furnaces in blast, and the like.
These series are very plentiful for the postwar period, fairly plentiful
from about 1870, and our stock of them for the more remote past increases
steadily. Many are available monthly or even weekly, and all may be
said to mean, approximately at least, a definite real thing that it requires
no theory to understand. Qualification is necessary, however. Some
of the modern and much of the older material is untrustworthy or at least
inexact, such as earlier estimates of crops or of pig iron produced or, in
times and situations in which smuggling is a factor of importance,
figures of imports. Some data raise questions of units of measurement
and comparability.
Changes in quality or territory and so on cast doubt on the value of
many series. Changes in sources and methods of compilation introduce
spurious breaks and fluctuations. The theorist's questions — what is a
commodity,1 a factor of production, a country? — acquire ominous sig-
nificance for the most practical purposes. In such cases as that of
building permits, the meaning itself becomes doubtful. We not only
open up obvious sources of serious error, but also obscure essential
features of cyclical movements, if from figures measuring one stage of the
career of, for example, a raw material, we draw conclusions about another
stage — from iron, conclusions about industrial equipment; from wheat,
conclusions about bread; or from imports or exports, conclusions about
production. Finally, the changing significance and efficiency of indi-
1 If we call motorcars a commodity, we are immediately, as in the case of prices, faced
by an index problem. If we restrict the concept to model X of the firm Yt the material
becomes unmanageable. The factor labor brings out the difficulty best of all. A country
like the United States, or even France, splits into sectors much more different in character
than Venezuela is from Colombia.
483
484 BUSINESS CYCLES
vidual commodities in consumers' budgets and in the pattern of pro-
duction, itself a most important feature of the results of the cyclical
process,1 invites erroneous interpretations if the series is not studied in
relation to the history of its industry and technology, which alone gives
the key to its meaning. Another research program unfolds itself, quite
beyond the means of the individual worker.
If we form composites from groups of related commodities, such as
foods, furniture, equipment, textiles, and the like, we meet, only with
more inescapable clearness, a problem that differs but in degree from
the problem which is implied in speaking of a quantity of coal in general
or coffee in general. A heap of woolen and cotton fabrics is possibly
less wanting in meaning than a heap of iron and strawberries would be.
In the most favorable cases, notably in cases of complementarity and
rivalry in fixed proportions, but also beyond these, even exact theoretical
meaning may be attributed to the composite. We shall not, however,
enter into these problems but, simply relying on the common sense of the
thing, confine ourselves to the remark that such composites derive addi-
tional justification if their constituents are, owing to their place in the
economic organism, all exposed to similar external and internal influences.
As with group prices, it is necessary to bear in mind that composites may
seriously obscure what precisely is the essential fact about the cyclical
process of economic evolution.
But the concept of total output lacks similar meaning or justification.
The fact that it has gained citizenship and is "recognized" is no comfort,
for this recognition has been extended to it quite uncritically. The prob-
lem is much more doubtful than the corresponding one in the case of
prices. There, we have at least been able to discover and define an
economic magnitude that supplied the meaning of an index of the price
level. Here, no such thing actually exists. However useful for many
purposes, total output is a figment which, unlike the price level, would
not as such exist at all, were there no statisticians to create it. We seem
indeed to be faced by a meaningless heap.
Three ways are open to us in order to overcome this difficulty, although
none of them can be considered as entirely satisfactory. The first has the
1 In particular, products require as time goes on, less and less of raw materials, such
as coal, steel, and sugar beet, per unit. For many purposes, series should be corrected for
this. It certainly vitiates the implications of rates of increase in the production of those
materials. The use of scrap is an instance of difficulties of another type. Improving
quality, still another striking feature of economic evolution, works the same way. It also
largely escapes us. We gather from historical indications that even the oldest and most
ordinary articles of consumption, such as meat and wine, are quite different now from what
they were even 100 years ago. But in most cases we have no means of measuring the
change. This really casts doubt on the possibility and meaning of any statement that
turns on any but the most outstanding features of our graphs.
PHYSICAL QUANTITIES. EMPLOYMENT 485
merit of simplicity. It was mentioned in the first chapter and consists in
fastening upon some series the items of which measure something that
may be taken to indicate the variations in productive activity — cyclical
variations in particular. All quantity series have some "systematic"
significance, but those we have decided to call systematic have much more
of it than others. Their limitations are, however, serious. Employ-
ment, even if statistics for our period were better than they are, would be
seriously misleading for periods longer than, say, a Juglar, often for even
much shorter periods: in good theory, it can never be assumed to be
proportional to output (see Sec. D). Tonnage cleared, not valueless for
Great Britain, is open to objections, not only because of the changing
significance of the unit and the changing importance of international trade,
but because it puts a ton of coal equal to a ton of gramophone records.
Similar difficulties surround the use of figures of freight carried by rail-
roads and, for example, horsepower installed. Pig-iron consumption,
though of course not above criticism, does astonishingly well for prewar
times, if tested in the light of our historical knowledge of cyclical fluctua-
tions.1 We will therefore immediately present the series and also the
series of rates of change of pig-iron consumption, Charts XII and XIII.
The trends in Chart XII, of course, are the trends of pig-iron consump-
tion and of nothing else. They do not give the trends of total output,
since the equipment into which pig iron enters increased faster than other
elements of total output, nor do they give even the trend of total "real"
investment. Moreover, pig iron cannot be trusted to give the cyclical
variation of total output, since cyclical variations in consumers' goods'
production are not proportional to, or synchronous with, cyclical varia-
tions in the production of equipment. But the fluctuations in the produc-
tion of equipment (steel did not in our period enter so largely into
consumers' goods as it does now) in all three cycles it gives very well.2
This is particularly true of the Juglar, which it is hardly possible to mark
off so neatly in any other series — the reader should mark it off here and
use the result for reference. The three curves in a sense interpret
each other. The American curve, otherwise the most satisfactory one
and that which shows the firmest step, presents a problem for 1875 to 1885
1 Pig iron held its own fairly well until about 20 years ago, though the process of ousting
it had set in by the sixties. Later, castings were quickly going out and weldings became
of much Jess weight. Lighter alloys also came in, and other metals gained relatively,
particularly aluminum. The increasing use of scrap has been mentioned. On the other
hand, iron and steel gained ground at the expense of timber and other materials. As
has been mentioned, consumption of pig iron has been used as the backbone of the analysis
of cycles by Professor Spiethoff.
2 There is some interest in noting that England's pig-iron consumption hardly took part
at all in the Kondratieff prosperity after 1897 — another indication of the fact that the
entrepreneurial impetus in English industry began to slacken about that time.
BUSINESS CYCLES
V
\
fft
58 L
IS «
I I I I I I
*->
ex
M
I
PHYSICAL QUANTITIES. EMPLOYMENT
487
that we know how to solve as soon as we glance at the English a^nd
German lines, the behavior of which is in turn cleared up by the American
line in the segment 1885 to 1892. The differences between the three lines
in average amplitudes and general character are very revealing of the
pulse of economic evolution in the three countries. The Kitchins, so the
writer thinks, show with sufficient clearness on Chart XIII. No doubt
some readers will not agree. To probable objections it can only be replied
that it seems a mistake to refuse to consider as a distinct class fluctuations
which are so similar in nature and so clearly marked, only because they do
Per
Cent
+20
UNITED STATES
UNITED KINGDOM
A /
A
-20
+20
0
-20
+20
0
-20
1855 1860 1865 1870 1875 1880 1885 1890 1895 1900 1905 1910
CHART XIII. — Rate of percentage change of pig iron consumption (see Appendix, p. 1056).
not display a regularity that simply is not in the phenomenon. Chart
XIV is presented in order to indicate how far other series of more than
average systematic significance could render the same service.
We may, in the second place, define variations in the "physical
volume" of any composite to mean variations in its dollar volume cor-
rected for changes in price level. The result, it should be noted, is still a
value, and not a true physical quantity, as we might expect according to
those definitions of price level which make it a sort of average price.
Our level is not a price but a pure number. Hence "deflated dollar vol-
ume" is simply a value figure from the variations of which the effects of
the variations of the monetary parameter have been removed. However,
V
..A
GERMANY
A
488
BUSINESS CYCLES
COAL PRODUCTION
/
COTTON »RODUCTIONy
/
V
RAILWAY FREIGHT TON-MILES
f
/BUILDING PERM |rS
COTTON CONSUMPTION
Vw'
1840 45 I860 55 1860 65 1870 75 1880 85 1890 95 1900 05 1910 13
CHART XIV.— United States (see Appendix, p. 1056).
PHYSICAL QUANTITIES. EMPLOYMENT 489
this has definite meaning, and these corrected figures may for some pur-
poses be treated as if they were physical quantities. It should be observed
that, in order to effect this, it is necessary to deflate by an index which
approximately presents our level concept, and not by any other — for
example, an index specially constructed to include only the prices of those
commodities that enter into the given composite, with weights that cor-
respond to the relative importance of those commodities in the composite.
We are not concerned with the question whether the latter proceeding has
any meaning of its own. All that matters is that this meaning is a dif-
ferent one.
The money value of Total Output (unfortunately along with other
items) is reflected in Outside Clearings, the value of Total Output of
Consumers' Goods, in the Sum Total of Private Incomes minus savings
and taxes paid out of income. Where we have these two figures, or at any
rate one of them, we might hence try to solve the problem in hand
by deflating. The reasons why we do not primarily rely on this possibility
will be apparent later when we discuss the Clearings and Income series.
Insuperable theoretical scruple is not one of them. But although we make
bold to use as a level series what we know to be something else, and
although we may be equally bold in the case of Clearings and Incomes
taken by themselves, we hesitate to cumulate errors by combining the
two, even apart from various other objections on the score of statistical
method.1
A third method of arriving at a single figure indicative of variations
of total physical quantities of commodities produced or consumed (services
and voluntary leisure ought, strictly speaking, to be included), consists in
constructing an index from individual quantity series. Everybody
recognizes that quantities produced, consumed, and in stock, ought to be
separately combined and that these separate indices should be confronted,
but everybody puts up with a combination of everything that is to be had.
The leading contributors all have a preference for weighting individual
relatives, or relatives adjusted for some trend and for seasonal variations,
with the American Value Added by Manufacture or the English Net
(Value) Product, where products of different stages of a productive process
are included. But, in cases in which these values are not available, some
fall back upon other criteria of relative importance, such as workmen
employed, pay rolls, horsepower installed, and so on — none of which they
1 Mr. Snyder's "A New Clearings Index of Business for Fifty Years." Journal of the
American Statistical Association, 1924 p. 829, does precisely this, except that he deflates by
his General Price Level, which, for the case in hand, is an index specially constructed to
correspond to the various components of the Clearings figure. The above argument is not
intended to imply adverse criticism of this interesting experiment.
490 BUSINESS CYCLES
would defend on theoretical grounds.1 The formal theory of such an
index is easy to derive. All we need to do is to change the Laspeyres
formula for the price level (see Chap. VII, Sec. B)LP = 2piq0/2p0q0 into
the obviously equivalent expression Lp = — —^ - , where the values
^PoQo
p0<lo are, in the numerator, now used as weights of price relatives, and to
*
form the analogous expression for quantities, viz. Lq
which means quantity relatives weighted by values, though not by added
values. But putting, as we did before in the case of prices, q% = q0 + dqot
and dropping subscripts, we get
1 There are available for the postwar time and particularly for this country indices of
this kind which excel in careful construction and technical perfection. Given the limitations
imposed by the material and granted the principles upon which their construction proceeds,
they probably do as much as can fairly be asked in all such matters as assembling and
criticizing the material, and correcting for seasonal variations and differences in working
days. The pioneer work of Mr. E. E. Day and the Harvard Society, its continuation and
amplification by the Federal Reserve Board, Prof. Warren Persons' and Mr. Snyder's
contributions, the indices of the Standard Statistics Company and of Mr. Leong, the
English index of Mr. Rowe and the London and Cambridge Economic service, and the
German indices of the Institut filr Konjunkturforschung must be specially and gratefully
mentioned. In particular, the distinction, fundamentally important as we shall presently
see, between equipment goods, other producers', and consumers' goods, has been fully
attended to. For the prewar time we are less well off, of course. There is, again, Mr.
E. E. Day's work and the work of Professor Warren Persons. Earlier pioneers ought to be
mentioned, notably E. Leonard: Index of Changes in Extractive Industries (Publication of
the American Statistical Society for September 1918) and, although their investigation
bore on trade rather than production, Professors E. W. A. Kemmerer and Irving Fisher.
The subject is also indebted to Professor W. J. King, Mr. W. W. Stewart (see the latter's
Index of Production, American Economic Review for March 1921) and Mr. W. Thomas.
Mr. Snyder extended his researches back within his work on his index of trade. Professors
Warren and Pearson have made an extremely valuable contribution as to basic and agricul-
tural production. The Index of General Business of the American Telephone and Tele-
graph Company has, in 1922, been reduced to an exclusively physical index. Colonel
Ayres' index contains also prices and values from 1790 to 1855, but is exclusively physical
from 1855 to 1901, with, however, such weighting as would make it lean on the Day-Thomas
index used by him for 1901 to 1919. For England there were only various indicators and
individual bits of information (for the early part of our period, largely contained in Porter
and Baxter), but we may now use the index compiled by Mr. Hoffmann of the Kiel Institut
ftir Seeverkehr und Weltwirtschaft, which, however, had to work with material not admit-
ting of any rational system of weighting. It goes back to 1713 but we use it only from 1785,
because before that date component series are too few. Mr. Snyder's index of English
production should also be mentioned. For Germany, we use the Index of the Institut flir
Konjunkturforschung, compiled by Mr. Wagenfiihr, who has had to substitute, in some
cases, for figures of production or consumption, figures of net import, railway transporta-
tion, or even money value.
PHYSICAL QUANTITIES. EMPLOYMENT 491
that is to say, an index of that part of the actual relative variation in
expenditure, (E + dE)/E, which as we have put it before, is balanced as
to its effects on the price level by changes in physical quantities and by
which expenditure would have had to change in order to keep the price
level constant.
Hence we have again a value figure from which the effects of changes
in the monetary parameter have been removed. The method is thus
seen to be but a variant of the method of deflating values. It aims at
what is essentially the same thing. However, it not only avoids the
difficulties that arise from the nature of the clearings figure and the
cumulation of errors incident to the process of deflating, but is theoretically
superior to the latter, in that it follows logically, and derives its meaning,
from the equation which embodies the theory of the price level and can,
owing to this fact, never give absurd results.1 If we recall how in the
fourth chapter we defined the effects of innovation on the depth and
breadth of the stream of commodities, we see immediately in what sense
such an index implements our propositions about variation of output in
the course of business cycles. It does not measure physical output in the
literal sense, but it docs measure physical output as transformed by the
introduction of an economic dimension. An argument would be in order
here for taking account of the structure of the world of commodities by
the choice of commodities to be included. But since it would, mutatis
mutandis, only repeat what has been said on this question in our discus-
sion of the level index, we will not stay to develop it.
B. The Analysis of the Trend in Total Industrial Output. — Quantities
produced or consumed are, unlike prices, rates per time element. Their
variations in the course of cycles are part of the primary as well as of
the secondary phenomena and, within both these categories, consequential
with respect to some, causal with respect to others — the most important
approach to a primary causal role being the relation of actual or expected
release of new products to the turn from prosperity into recession. An
index of Total Output makes a synthetic, systematic, and cyclical trend
series.
To verify expectation as to the presence of a trend, it is only necessary
to look at the quantity lines in charts V, VI, and VII. Moreover, Chart
1 It should be noticed that the Harvard Index of Physical Volume, which is in essence
such a corrected value, in the short run goes very well with Freight Cars Loaded, if both
are adjusted for business days per month and for seasonal variations, and treated according
to the Harvard method. See W. M. Persons, Car Loadings as an Index of Trade Volumes,
Review of Economic Statistics for October 1926. The same is true for Germany, see W.
Teubert, Der Gliterverkehr, Sonderheft 5 of Vierteljahrahefte zur Konjunkturforschung, 1928.
492
BUSINESS CYCLES
XV presents an international comparison that is of some interest,
although, owing to differences in materials and methods it must not be
relied on too implicitly. Chart XVI gives the same series as transformed
by the operation of empirical differentiation and, by the absence of any
marked trend, teaches us something about the character of the trend
present in the original figures. Of course, this does not mean that with
other data the same result would emerge, or that any trust can be placed
1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910
CHART XV. — Industrial production (see Appendix, p. 1056).
in the gradient of the particular logarithmic straight line which an investi-
gator's material may yield. Some authors who speak of a "compound
interest law of growth" seem, like the queen in the play, to protest too
much. Still more treacherous and pregnant with danger of speculative
temerity may be the application of Verhulst's formula, y — , _t , , >
which was intended (1888) to represent certain features of organic or of
similar types of growth.1 Even perfect fit in the least square sense would
1 Compare, for instance, Wittstein's formula in Gesetz der menschlichen Sterblichkeit,
1883. Verhulst's formula is used (slightly generalized) by A. J. Lotka. The form 6f , •
is known in this country as the Pearl-Reed curve.
PHYSICAL QUANTITIES. EMPLOYMENT
493
y-
iu-u
I 11
- £
8 I
si
«•»
a &
"8
CW a,
if
«g
i
8 3
^ ol^
494 BUSINESS CYCLES
not prove anything. We are, however, on somewhat safer ground when
applying such expressions to the behavior in time of quantities of indi-
vidual commodities.
The broad fact of great steadiness in long-time increase nevertheless
remains, both in the sense of rough constancy of the gradient of the trend
and in the sense of what, merely by way of formulating a visual impression,
we may term the general dominance of trend over fluctuations. As a rule,
even strong depressions such as the English one of 1825 could hardly be
identified as such from the output graph alone. In no country does 1873
look very catastrophic. In America, 1884 produced almost no fall at all.
The crisis of the early nineties shows, for Germany, by only an incon-
siderable dent. In the long English series it happens only twice that
absolute fall outlasts two years. In the case of Germany, this occurred
only in 1868, 1869, and 1870; in America also but once. One of the most
important tests of an analytic model of the cyclical process of economic
evolution is whether or not it enables us to understand that steadiness
in both senses (such as it is).
What we see, is, of course, only a descriptive trend. In the same sense
as in the case of the price level we interpret it as a result trend, blurred and
deflected by outside disturbances which, particularly if a trend line be
fitted by least squares, may acquire lasting influence, even if passing by
nature. And there is also the effect of growth. As has been pointed
out before, no satisfactory method is known to the writer of eliminating
these influences, and what follows must be read with due regard to the
qualifications that this disability implies. The result trend itself must
be explained from the behavior of output in the cycles that generate it.
Before taking up this subject, we will digress, in order to comment on
some questions which, though they do not strictly lie on our path, yet
come too near to be passed unnoticed.
It is reasonable to believe that if our analysis enabled us to isolate
and measure the result trend in output plus the effect of growth, instead of
merely teaching us to recognize their presence in our series, the output
curves Would turn out to be steeper than the descriptive trend, because
practically all the other factors which influence the latter are of the nature
of injuries inflicted upon the economic organism. This way of expressing
a patent historical fact should not be understood to imply adverse judg-
ment about what, from our standpoint and for our present purpose, would
have to be included among injuries. Many measures of social better-
ment that most people heartily approve of and many nationalistic policies
that command fervent allegiance come within the meaning of that term,
simply because they prevent the economic machine from working accord-
ing to its design, but nothing is further from the writer's mind than to
hold that it "should" be allowed to do so, or that everybody would be
PHYSICAL QUANTITIES. EMPLOYMENT 495
happier if it were, or that abundance is the criterion by which to define
welfare or to judge a civilization. Nevertheless, the descriptive trend of
output is both in itself and because of the idea, however vague, which it
gives us about what the combined effects of evolution and growth would
be in the absence of other factors, for a purely economic analysis, a fact
of importance. In a sense it is the most important fact about the eco-
nomics of capitalist society as distinguished from its culture and the type
of man it creates. In beholding it we have before us a measure of the
actual economic result of humanity's great experiment with private busi-
ness and the acquisitive principle.
Whether the gradient we observe is only, as we implied above, what is
left of an even greater possibility or whether the injuries referred to are the
outcome of the capitalist process itself and, hence, to be recorded against
it and inseparable from it, need not detain us here i1 this is a question that
cannot be dealt with without entering into the theory of the nature and
behavior of social classes. Nor does it concern us how capitalist perform-
ance compares with the performance we could expect, or could have
expected at any time, from alternative institutional arrangements.
Theories that seem convincing and are sympathetic to some have been
offered, in order to make out a case for the proposition that the economic
engine characterized by private enterprise necessarily turns out the
maximum of output which it is, or was, possible to produce within the
framework of natural, technological and political data. Other theories
that seem convincing and are sympathetic to others have been offered
in support of the proposition that it is of the essence of the system of
private enterprise to produce less than could be produced within the same
circumstances. There is little to choose, as regards looseness of logic and
carelessness about facts, between the two types of argument. We will
confine ourselves to noticing the following points.
First, from the indubitable fact that the capitalist arrangement tends
to draw the best brains into business pursuits and to spur them to do their
utmost, it does not follow that these efforts must necessarily issue in maxi-
mum production, although that fact remains relevant to the argument.
Second, for perfectly competitive conditions it can be proved, with
but unimportant exceptions, that states of perfect equilibrium will
in fact be characterized by maximum output. This proposition is
not made entirely valueless or tautological by all that has to be assumed
1 We met that question before in our discussion of the concept of external factors and
repeatedly in our historical sketch. We take here the same standpoint as we did there:
there is obvious sense, we admit, in saying that if a man takes to drink because of ill-health,
the effects of the alcohol are to be included in a full description of the working of his organ-
ism. But they are still due to a distinguishable factor which for many purposes should be
dealt with separately. We shall not leave this standpoint except for a short passage in
Chap. XIV.
496 BUSINESS CYCLES
in the process of proving it. Refinements of analysis often render what
amounts to a disservice in such matters by unduly emphasizing qualifica-
tions of little practical significance. But the analogous proposition
can similarly be proved for the case of a socialist organization.
Third, any argument to the contrary that runs on the lines of economic
principle at all must mainly rest on imperfections of either competition or
equilibrium. These, however, yield so strong and so universally recog-
nized a case against the pretended efficiency of the capitalist machine that
it becomes necessary to recall, on the one hand, that a system in which
imperfect competition prevails will, contrary to established opinion,
produce in very many, perhaps in most cases, results similar to those which
could be expected from perfect competition and, on the other hand, that
even if a system consistently turned out less than its optimum quantity,
this would not in itself constitute disproof of optimal performance over
time. 1 Whoever wishes to hold, either that imperfect competition entirely
inverts the working of the system, or that failure to produce instantaneous
maxima spells failure to evolve along a path of maximum production
in the sense alluded to, will therefore have to embark upon a much more
arduous analysis than is implied either in reasoning from a few assump-
tions or in analyzing individual facts without relating them to the whole of
the economic process. He will be on much safer ground, however, if he
starts his attack from other than purely economic standpoints.
Fourth, whoever wishes to hold that capitalist performance in fact
lives up to the possibilities which the capitalist system itself creates, and
that either the descriptive trend, or what one may conceive the true
result trend plus growth to be, really represents the possible maximum
output over time, is of course under the same necessity. Such a venture,
to which theory can only contribute a technique of inference, is in the
present state of our factual knowledge probably beyond the means of the
individual worker.2
1 For any dynamic system (economic or other) may work in such a manner that it
produces its optimum (with reference to any criterion) over time, only at the price of never
reaching it at any point of time. This is not more paradoxical than, for instance, the fact
that producing at minimum cost over time may, in the case of lumpy factors, imply never
producing at minimum cost on any of the short-time cost curves because another "method"
becomes more advantageous before that point is reached. The proposition may further be
illustrated by that kind of excess capacity which is an unavoidable incident of quick
"progress." The first of the two propositions contained in the above sentence may be
illustrated by that case of monopolistic competition in which there is a great number of
actual and possible varieties of a commodity that are highly substitutable for each other.
Part of what has been said on imperfect competition in the second chapter will help to
establish the point.
* Many engineers and efficiency experts, particularly of the technocratic type, are not
of this opinion and find it extremely easy to collect overwhelming evidence of underutiliza-
tion of resources. But this only proves that they have not even seen the problem.
PHYSICAL QUANTITIES. EMPLOYMENT 497
But whether or not the descriptive trend traces out a path of maximum
production — be this maximum relative to capitalist conditions or to more
general ones — the fact remains in any case that such performance as that
trend embodies is not only historically associated with that concentration
of human efforts on private economic ends, with — perfectly or imperfectly
— competitive private enterprise, with intensive saving, in short with
"profit economy," but also either wholly or to a considerable extent
historically dependent upon it : so much is clear from our analysis and so
much would — witness the Communist Manifesto-^have been admitted by
Marx himself, though not by many of his followers. Since extrapolation
of the descriptive trend yields the result that the profit economy would, if
allowed to work on, do away, in a not distant future,1 with anything that
can according to present standards be called poverty, it is of some interest
to inquire whether there is any warrant for such an extrapolation. This
raises the question of the phenomenon of retardation in rates of increase,
which some students believe our physical series display.2
Within the range of problems we are envisaging at the moment, the
question carries meaning only for total output or total output of con-
sumers' goods. For it is obvious that even if we disregard the phe-
nomenon, fundamental to economic evolution, of displacement of old
commodities by new ones, no industry can go on expanding output at the
rate of its innovation stage. Each reaches maturity in the sense that it
finds its place in the economic organism and the amount of output beyond
which it cannot profitably go, unless that amount be increased by some
further innovation within it or in some "complementary" industry and
by the general effects of increasing wealth (which, however, may also be
negative) and of Growth. But it is equally obvious that by itself neither
absolute rate of increase nor rate of increase per head of population
corrected for age distribution, will give us the information required, for,
to exemplify by a strong case, absolute decline in population may so
reduce the wants of the community that even strong and persistent
progress in efficiency of the productive engine could be powerless to
balance the effect on total output, while such decline may produce rising
per capita figures without any "progress" in our sense, possibly even with
a negative one. Finally, we should take account of variations in the
amount of voluntary leisure, which may be and undoubtedly often is
one form of taking increased real income in the Fetter-Fisher sense:
1 For the United States, we find, assuming the population to reach the figure of about
160 millions in 1978 and production to increase at 3 per cent compound interest, that income
per head would in that year be about $2,800, purchasing power of 1928, when income per
head was about $700.
2 The idea is an old one and has taken several different forms. But we will mention
only Professor S. Kuznets, Secular Movements in Production and Prices, and Mr. A. F.
Burns, Production Trends in the United States since 1870.
498 BUSINESS CYCLES
in this sense the reduction of working hours is one of the most significant
"products" of economic evolution. All this illustrates the difficulties of
speaking of retardation if the object is to measure capitalist performance
with a view to deciding whether or not there are "hitches " inherent to the
working of the profit economy that tend to thwart mankind's economic
possibilities. Further illustration is afforded by the following examples
of actual or possible causes of retardation.1
First, we will restate statistical reasons, all of them previously men-
tioned. As in the case of indices of price level, we have to accept the
fact that any index of total or of consumers' goods' output will in general
be influenced by individual and group developments. In the case of the
production index, this spells downward bias: new commodities and fin-
ished commodities are inadequately represented, services (and voluntary
leisure), not at all, while commodities that are in the process of being dis-
placed loom large. Improvements in quality, particularly in durability,
largely escape, and they may in many cases be increasingly important.
The quantities of raw materials and semifinished goods which form the bulk
of what enters into indices of output are, in the course of technological
progress and because of the spreading use of waste or scrap, made up into
increasing quantities of finished goods, so that observed retardation may
be spurious and even indicative of its very opposite. Finally, the period
1897 to 1913 covers the prosperity phase of a Kondratieff — a fact that
would, according to our theory, be sufficient to produce an impression of
retardation if comparison is with the years of Kondratieff depression and
recovery, 1870 to 1896.
Second, there are what we may term sociological reasons. Some of
them — which may account for some slackening in entrepreneurial effort
and cognate phenomena — have been alluded to in the third chapter and
will cross our way again in our discussion of the postwar period. But it
has also been pointed out that, in part at least, economic evolution grows
increasingly independent of the typically capitalistic pattern of cultural
values and motives. In any case, the writer believes it to be safe to
discard this group of factors in an interpretation of American and German
developments during prewar times, although in the case of England we
have seen reasons for some doubt about this.
1 The reader should compare Chap. IV of Mr. Burns's book mentioned in the preceding
footnote for a suggestive discussion. He will find that statistical difficulties, formidable as
they are, are secondary to the difficulties of interpretation which of course differ widely
according to the purpose in hand and can be expected to yield only to most careful analysis
of industrial processes in detail. Inferences as to decreasing returns or any slackening of
rate of progress in efficiency (as measured by service) may easily prove misleading. As to
the immediate statistical result of Mr. Burns's investigation, the writer agrees with the
views expressed by Professor Mitchell in the preface to that book. Compare also Pro-
fessor Crum's review in Quarterly Journal of Economics for August 1934.
PHYSICAL QUANTITIES. EMPLOYMENT 499
Third, we have the economic reasons. What eventually may turn
out to be the most fundamental of them and also account for some of the
sociological ones — increasing satisfaction of wants — can hardly have been
operative in our period. For it takes more for it to assert itself than the
mere fact of declining marginal utility of income, and of the many subtle
elements that enter into this problem none can be identified with any
certainty unless indeed we interpret certain cases of the struggle for
shorter hours in this sense. Nor can decreasing returns, in the sense of
increasing difficulty of procuring food stuffs and raw materials, have
played a major role for the period as a whole. Any Kondratieff pros-
perity will of course tend to display — up to a considerable level of general
wealth — rising prices in both those groups, and this each time deceives
many economists into believing in a secular law of decreasing returns.
But it is a temporary phenomenon or, at all events, it has been a tem-
porary phenomenon during our period.
Again, it is precisely the increasing abundance of foodstuffs and raw
materials within the period that has by some economists been made the
basis of an argument that would, if true, tend to establish retardation
in a special sense and at the same time explain the failure to come true of
Marxian predictions about the mass misery which was to result from the
working of the capitalist mechanism. It is held that there would have
been retardation either from decreasing returns or from the nature of
the profit economy but for the fact that the opening up of new countries
temporarily put decreasing returns out of operation, and that it is bound
to set in as soon as this unique opportunity is exhausted. We have
noticed, in Chap. I, that it is not correct to consider the actual opening
up — as distinguished from the discovery — of new countries as a factor
external to the economic system. The processes of exploiting a new
country are but a type of innovation and alter the data of the economic
process in no other way than do other innovations. Hence, their effects
are properly included in a measurement of capitalist performance. Nor
does any prognosis of retardation follow, for it is only one of many types of
innovations — and one only of the many types of innovation in the fields
of foodstuffs and raw materials — the possibilities of which thus become
exhausted. The possibilities of each individual type do so, of course,
but nothing can be concluded therefrom for innovation in general.
Most of us seem here to commit a mistake in handling the concept of
decreasing returns. In its proper sense it applies to given production
functions and generally stationary conditions plus Growth only. In
order to make it relevant to any forecast about the future course of
production, it must be used, as indeed it has been used already by
Ricardo, in a different sense, namely in the sense that action of the "static
law" of decreasing return will indeed be interrupted by innovation but
500 BUSINESS CYCLES
that the latter will be powerless to compensate its effects in the long run
— that, as it were, there is a law of decreasing returns from successive
innovations. And in this sense the statement is entirely unwarranted.
Owing to their unpredictability, future innovations may be less or more
conducive to increase of means of satisfying wants than were previous ones.
"The great things may be done" but they may equally well be still to
come : while the earth can be surveyed and better opportunities (relative
to each given state of technique) for its exploitation jnay be taken up
first, so that inferior ones only are left for the future, the world of possible
innovation cannot be mapped out.1 Postwar agricultural history testifies
conclusively on this point.
C. The Cyclical Behavior of the Physical Volume of Production. —
It will be recalled, first, that our theoretical expectation as to the behavior
of total output was the net result of our theoretical expectations as to
the behavior of two constituent groups centering respectively in pro-
ducers' goods and the commodities that enter into the budgets of house-
holds and, second, that it changed with successive approximations.
What that net result will be in the case of any given index cannot be pre-
dicted unless we know what commodities are included and how it is'
constructed. This applies with greater force here than in the case of the
price level, because total output is not a definite real thing. The com-
parative steadiness (in the sense, say, of "good" fit to the material of an
expression of the form y — ax1) of our output series is not in strict theory
primarily due to steadiness of its individual constituents, nor simply
to the smoothing effect on random fluctuations in constituents of combina-
tion into an aggregate (though this effect will also be present, of course),
but to different rhythms in the two groups of constituents. These
rhythms, already somewhat toned down in comprehensive composites
of producers' and consumers' goods,2 are further reduced in amplitude
by a number of factors mentioned in Chap. IV, which we will presently
restate.
Our expectation that total output will increase through all phases of
the cycle, "deep" depression alone excepted, is derived as follows — the
exception hardly ever extending over the whole of the depressive phase,
1 Professor Kuznets, op. cit.t in fact, assumes the existence of such a law of decreasing
returns of economic progress, and has tried to establish it by analyzing the effects of
successive innovations in technology. The partial success of this analysis is, however,
merely due to the fact, noticed above, that in every industry innovation tends to exhaust
itself for the time being and to taper off into comparatively unimportant improvements of
the induced or completing type.
2 For obvious reasons, durable consumers' goods will display a tendency to conform to
the behavior of equipment goods. We shall repeatedly refer to this fact, which tends to
accentuate fluctuations. So does the practice, widely observable, of putting off the replace-
ment of equipment in depression or even concentrating it in prosperity.
PHYSICAL QUANTITIES. EMPLOYMENT 501
since it is due to panics and vicious spirals, which as a rule do not last
more than one year.1 As for equipment industries and their satellites,
expectation from our model does not differ from accepted opinion and
only provides explanation of a fact universally admitted and sometimes
even overstressed.2 Quantity taken should increase in positive phases
and decrease, or increase less, in negative phases. Quantity produced
may be, and undoubtedly often is, particularly in the course of the shorter
cycles, interfered with by exports and imports and other circumstances,
but should still move substantially the same way. Of course it is neces-
sary to take account first of the interference of cycles with each other;
second, of the fact that if entrepreneurs9 demand for equipment slackens
— or, in strict theory, is absent in three of the four phases — the demand
for equipment that is induced by entrepreneurial activity not only does
not cease, but actually concentrates in recession and revival; and, third,
of Growth, particularly in the Kondratieff. Inasmuch as we may take
pig iron as a representative of the equipment group, our chart of pig-iron
consumption may reasonably be said to verify that expectation. Also,
the chart of rates of change should now be compared with the chart of
rates of change of total output.
We shall expect wholesale prices3 to precede the output of pig iron
or, for that matter, of equipment goods — at least, in the shorter cycles —
not only because transactions precede production and because of the
effect of speculative anticipation, but also because wholesale prices will in
1 As in the case of the price level, we shall associate them with the two short cycles,
because in the long and gentle sweep of the Kondratieff they would hardly arise. This
does not mean that the Kondratieff has nothing to do with them: conditions of the under-
lying cycles always influence events in the course of the superimposed ones.
2 We may illustrate that received opinion, first effectively preached by Tugan-Baranow-
sky, by a quotation from Marcel Lenoir, Etudes sur la Formation et le Mouvement des
Prix, 1913 p. 77, of the implications of which we, of course, do not approve: "la cause pro-
fonde des fluctuations periodiques de la vie economique . . . est la demande intermittante
de capitaux fixes necessaires, de temps en temps, pour la refection et 1' augmentation de
1'outillage economique." If refection and augmentation were all, it would not be easy to see
why demand for "fixed capital" should be intermittent. But there is almost universal
agreement about the fact.
3 We shall not be so sure about the price of pig iron itself. In strict theory, this price
should not increase in revival, since the demand for new equipment for new purposes does
not set in until the end of that phase. But both replacement demand and demand from
investment opportunity created by innovation in the previous prosperity will assert them-
selves in recovery. On the other hand, iron and steel, being cyclical industries and holding
a reserve of capacity, need not react by increasing their prices at first, i.e., in revival, or
even at the beginning of prosperity. The reader will find all this very complicated. If
thereby he means that it would be more pleasant if reality were simpler, the writer can only
heartily agree with him. If he thereby means to record an objection to our analytic
schema, the writer does not agree. He is unable to see any merit in a theory yielding simple
and clear-cut propositions which do not fit facts.
502 BUSINESS CYCLES
general have to make up for a fall in depression and hence display an
increase in revival. According to the method of "lags of maximum cor-
relation" (Review of Economic Statistics 1919, p. 184, et seq.) this, in fact,
is so. If indices of producers' goods consist wholly or mainly of equip-
ment goods, the same will hold true for them. Of course, no causal role
of prices or of the mechanism of money and credit follows from that.
But our expectation as to finished consumers' goods and their satel-
lites, which finally boiled down to expecting that their output should
increase more in recession and revival than in prosperity, runs counter
not only to public opinion, but also to the opinion of most students of
cycles. In its "pure" form, our model yields the expectation that in
prosperities the output of producers' goods should at first increase at the
expense of the output of consumers* goods. The latter should, even
absolutely, decline, and if the innovations in question be of the type of
railroad construction, no fully compensating increase of the former
might show for years. In general, however, new products will be released
as prosperity wears on, their impact being part of the mechanism that
eventually turns prosperity into recession. Hence decrease could show
only in a segment between recovery and the later stages of prosperity.
But the avalanche of consumers' goods and their satellites should come
in recession. Their output should continue to increase, barring panics
and spirals, in depression and display the strongest increase in recovery.
But this principle holds strictly true only if prosperity starts from
perfect equilibrium in perfect competition. This not being so in reality,
business will respond to brisk demand by almost immediate expansion of
total production and no decrease in consumers5 goods production need
occur in prosperity. The presence of cyclical industries acts especially
strongly in the same direction. Moreover, the facts that we refer to
under the heading of Growth must blur the picture. The population of
the United States increased from 1839 to 1915 at a compound interest
rate of 2.28 per cent per annum,1 and we may assume that for most of
that period no tendency toward diminishing returns exerted appreciable
effects. This alone would be sufficient to produce an increase of physical
production during prosperity phases and to wipe out the effects of any
tendency in the opposite direction such as could be expected from the
working of our mechanism. But while this tendency will therefore gener-
ally fail to show, dips due to the breakdown of the Secondary Wave and
to the other depressive factors will almost always show. They influence
both people's opinions about the nature of downgrades and the statistical
1 It should also be taken into account that all the time the average age of the population
was increasing. It was around seventeen years, 1790 to 1810 (white males only). It was
eighteen by 1820 (whites only), and twenty-three years by 1910 (all persons). See Census
of 1920, vol. II, p. 148.
PHYSICAL QUANTITIES. EMPLOYMENT 503
picture. It should be observed that they are likely to distort a short
cycle more than a long one, in the course of which there is time for things
to straighten themselves out into their true form. We shall, hence,
expect the facts to conform best to the above expectation in the case of
the Kondratieff, less so in the case of the Juglar, least of all in the case
of the Kitchin cycle. A spurious deviation from expectation should also
be kept in mind, in this as in all other cases. If there are four phases
and depression contains a span of actual fall of output, then revival will
be the phase in which what would happen in the downgrade of a two-phase
cycle will be most clearly in evidence. But most writers date cycles
from troughs, and to them it must of course seem as though we were
contradicting an obvious fact.
As stated above, we must in this case, as in all, read any statement
with reference to the presence of three cyclical movements that super-
impose themselves on, and interfere with, one another. Although this is
more difficult to take account of, because the span covered by material
at all reliable is so short, it thereby loses nothing of its importance.
Investigation into the details of each small wave would be necessary in
order to illustrate it. On the downgrade of a Kondratieff, the increase
in consumers' goods incident to its mechanism will tend to overshadow
any opposing tendency of opposite phases of the shorter cycles. But
this is not all. It must be taken into account also that, as in such a
downgrade everything is technically and commercially prepared for
expansion, this expansion is particularly likely to show under the influence
of increasing expenditure induced by prosperities of the shorter cycles,
while their depressions, notably if coinciding, are likely to spell panic and
paralysis of business and so to inhibit the release of the products of the
new industrial apparatus, amid general complaints about price wars,
cutthroat competition, and overproduction. On the upgrade — witness
the "hungry forties" and conditions around the turn of the century —
the sweep of the Kondratieff intensifies the tendencies of prosperity
phases of the shorter cycles and weakens those of their recessions and
revivals. An example showing how important it is to attend to that
phenomenon is afforded by the indubitable fact, noticed in the preceding
section, that from material covering 1870 to 1913 it is possible, for all
three countries and many others, to show the presence of a long-time
tendency toward declining rates of increase in the production of consumers'
goods. As mentioned before, this is simply due to the circumstance that
the period 1870-1897 happens to lie in what first was a depression and
then a recovery phase of a Kondratieff, while the years 1897 to 1913
cover the prosperity and only a few years of the recession phase of a new
Kondratieff.
504 BUSINESS CYCLES
Returning to charts XV and XVI, it seems reasonable to say that the
theory submitted is borne out by the behavior of total industrial produc-
tion, l inasmuch as the theory affords a rational explanation of the steadi-
ness of that behavior. For if this theory were not correct, it would be
extremely unlikely that longer periods which we identify as nonprosperities
should display on the whole at least as high a rate of increase as periods
of prosperity, and sometimes a higher one. And if alternative theories
were correct which associate increase of output primarilynvith prosperities
and decrease or smaller increase with negative phases, it would be equally
unlikely that this should not show at all in the Kondratieff, which so
strongly displays the cyclical variations in other series. Even indications
pointing to the effects to be expected from our pure model are not entirely
wanting. If the reader will draw straight lines through the intervals
1898 to 1913 and 1858 to 1897 on Chart XIV, he will find, as has been
mentioned already, that for all three countries the gradient of the first
line is smaller than that of the second. And Professor Mills' statement,
(Economic Tendencies, p. 244) to be noticed again in our discussion of the
postwar period, undoubtedly lends support to the opinion that in the
third Kondratieff output actually increased in recession at a higher rate
than in prosperity. Something like this seems also to be true of the
English index from 1820 (if, on account of the Napoleonic wars and their
aftermath it is permissible to begin with that year instead of some date
around 1800) to 1842, as compared with 1842 to 1858.
Comparison of the behavior of equipment and of consumers' goods,
presented on Chart XVII, brings out, first, some familiar features, notably
greater amplitude of fluctuations in the former. But a lead of equipment
goods is not particularly in evidence. There is no reason why it should
be, if we look at peaks and troughs — consumers' goods may even be
expected to recover more quickly from troughs, and this shows in several
instances. But even apart from this, lead and lag in the short run are
largely, from the standpoint of our process, a matter of chance or rather
of peculiarities of individual situations that have little fundamental
significance. Disregarding questions of lead and lag, however, the cyclical
variation in the relation between equipment and consumers' goods
shows clearly enough. It does so especially in the three lines at the
bottom of the chart, which are particularly useful for the purpose of
1 There is really no such thing as a theory of total output as such, in the full sense of the
word, for this would imply that we are able to represent it as an explicit and uniquely deter-
mined variable in function of some independent variables of the same, i.e., the systematic,
class — such as price level, interest rate, quantity of circulating medium. Our whole analy-
sis shows this to be impossible and any theory that attempts it, to be a sham. What we
mean above is simply theoretical expectation as to behavior of output withiu a process that
simultaneously shapes all the variables,
PHYSICAL QUANTITIES. EMPLOYMENT
505
studying Juglars. Only the English series, characteristically different
from the other two, behaves, if anything, contrary to expectation as* to
the Kondratieff, and shows hardly anything except a short irregular wave
PRODUCERS' GOODS
CONSUMERS' GOODS
PRODUCERS' GOODS
CONSUMERS' GOODS
RATIO
UNITED STATES
GREAT BRITAIN
UNITED STATES
JREAT BRITAIN
GERMANY
GREAT BRITAIN
I8GO 1865 1870 1875 1830 1885 1890 1895 1900 1905 1910 1913
CHART XVII. — Industrial equipment and consumers' goods (see Appendix, p. 1056).
and a trend easily explainable on Boehm-Bawerk's theory. It gives
some support to the view that during the period covered the stronger
cyclical impulses came to the English organism from other countries.
506
BUSINESS CYCLES
The German ratio is very interesting. It rises at first through what is
distinctly abnormal behavior of consumers' goods at the beginning of the
sixties, then goes on almost level (still marking Juglars, however) to
1875, jerks up after a dip in 1877 (which though a year later than in the
United States, is yet an early indication of the new Juglar), to continue
on the same level or slightly falling to 1894, after which the new Kon-
A >
Jk \ rf
/i \ \ / /
•• "
n
/BASIC PRODUCTION
1870 7S 1880 85 1890 95 1900 05 1910 13
CHART XVIII. — United States consumption and production (see Appendix, p. 1057).
dratieff is exceedingly well marked. The two levels might be associated
with depression and recovery phases of the second Kondratieff. The
American ratio is the most lively one. It marks the Juglars in the ratio
with a clearness that leaves little to desire and even — a point that the
reader should verify carefully because there is plenty of room for differ-
ence of opinion on this — the Kitchins.1 The rush to a new level, which
is not lost again, stands out well at the beginning of the third Kondratieff.
Chart XVIII illustrates this still better .
There is some point in comparing, before we leave the subject,
variations in output with variations in the price level. We recall at the
1 The Kitchins can be best observed, for all countries, on the chart of rates of changes.
PHYSICAL QUANTITIES. EMPLOYMENT 507
outset that positive association of the latter with the cyclical phases, or
at least three of them — rise in prosperity, fall in recession, and depression
— is normal in the sense that the working of the model tends to produce it,
but that we had to qualify that expectation in various ways — notably
with respect to the interference of cycles with each other — even within
the pure theory of the process. We have now to qualify further by taking
account of what has been said about the behavior of output. Never-
theless, the main contour is as required by our schema. The reader
should look at the three pulse charts and allow himself to be impressed
with the two great result trends of capitalist evolution, the downward
trend of prices and the upward trend of quantities, both products of the
cyclical process. Although we must not simply attribute the first to the
second — they are both elements in a process that contains many other
variables — we may take the picture to represent one of the outstanding
features of that process. At least, if business activity be measured by
industrial production, it would be patently absurd to associate it either
exclusively or primarily with periods of rising price level. We might
even use the fact of the fall (both the fall in each Kondratieff downgrade
and the resulting secular fall) of the level, together with the fact that
media of circulation continued to increase, for a short-cut method of
proving — although, to be sure, this would not be quite correct — the
proposition that the downgrade plus recovery is the time of harvesting
the fruits of innovation in the form of increased output and that given
the behavior of the other factors as it actually is — but this is a method of
reasoning that the writer always highly disapproves of, if he finds some-
body else using it — it is the release of those fruits that turns the price
level. Inspecting for instance the British chart, we find level and
output rising together, with what we might imagine would, without the
French War, have been fairly similar gradients up to 1800 (when output
turns within a short cycle before prices), where we suspect the turn of
prices would have come without the Napoleonic wars. Prices rise with
interruptions to the peak, which in our series comes in 1814; output
winds up steadily through its Kitchins. Then prices come down to 1820
(the Kitchin in output that corresponds to the hump in prices in 1817—
1818 looking not very different from its predecessors), and output con-
tinues to increase in utter disregard of all theoretical argument that has
been put forth to the contrary by so many economists. Then prices
sweep down — output sweeps up at an increasing rate to 1825 and turns
into the crisis of that year before prices. And after that, it continues to
increase — if anything, less during the period of rising prices (1849 to
1873) than it did in the preceding period of falling prices. It rises some-
what less than we should expect from 1873 to 1895, but still at about the
same rate, in spite of the prolonged fall in prices. And it rises somewhat
508
BUSINESS CYCLES
more than we should expect, with rising prices from 1895 to 1913, but
again at about the same rate.
In the case of Juglars, it happens much oftener that falling prices are
associated with dips in output — precisely in depressions or " crises " such
as, in England, in 1793 and 1794, then in 1884, 1885, 1886, 1892, and 1893,
in 1903 and 1904, and on other similar occasions, though by no means
consistently so. Prosperity phases repeatedly rise from falling levels,
Per Cent
+40 r
+30
+20
+10
WHOLESALE PRICES
V\l
UU4i
r/% «T / v *
i
-10
-20
-30
+30
+20
+ 10
U
\l
v
A
BANK CREDIT
ACCOUNTS
CIRCULATION
-10 -
+200-
+100-
BERLIN MARKET DISCOUNT RATE
-100-
+20-
INDUSTRIAL PRODUCTION
~IO!860 1865 1870 1875 1880 1885 1890 1895 1900 1905 1910 1913
CHART XIX. — German prewar pulse, rates of percentage change (see Appendix, p. 1057).
but only on Kondratieff downgrades, as, for example, the English pros-
perity of the twenties of the nineteenth century, preceding the crisis of
1825, or the English and German prosperities of the eighties, which,
however, proves nothing against the "normality" of increase of price
level in prosperities. There is no systematic association of recession
with falling output. But in the Kitchins there seems to be: although
instances to the contrary are not lacking, the evidence for positive associ-
ation of output with the Kitchin phases and, on the whole, also with price
level, supported as it is by theoretical considerations, is fairly strong.
PHYSICAL QUANTITIES. EMPLOYMENT 509
We present on Chart XIX the rates of change of the series we know
already from the German pulse chart. Covariation is much in evidence.
It seems to be approximately instantaneous, although rate of change of
output displays a tendency to precede.1
D. Employment of Labor. — This is the only case of employment of
productive resources with which we shall deal. Investigation of the
variations in the degree of utilization of natural resources constitutes
one of the unfulfilled desiderata of our program. As regards variations
in the degree of utilization of plant and equipment, we must, for lack of
the kind of information we need for our purpose, confine ourselves to
repeating that, the more successfully progressive a community is, the
more underutilization or excess capacity of plant and equipment it must,
other things being equal, display, both because rapid progress means
violent disturbance and correspondingly violent upsets, and because the
more rapid the advance, the greater the percentage of scmiobsolete and
obsolescent plant and machinery will be which is not yet definitely dis-
carded by firms or dropped from statistical lists. Material concerning
variations in the employment of labor is very scarce and defective for
the whole of the prewar period. For Germany we have indications only
— mainly from the nineties on — which may easily mislead. The decen-
nial United States census data are no good for our purpose. The
1 Professor Irving Fisher (Our Unstable Dollar and the So-Called Business Cycle,
Journal of the American Statistical Association for June, 1925) has found by the method of
lag distribution that maximum correlation between W. M. Persons' index of the volume
of trade and variation in the rate of change of the price level is obtained if we lag the former
by seven months. If any trust can be placed in this lag, which is, as it were, the center of
gravity of the "effects" of the change, and shows the maximum correlation of those indices
for the period from August 1915 to March 1928, it would be rather tempting to urge that
a lag of that length would, in fact, roughly correspond to the distance, in a perfectly regular
Kitchin cycle lying on a Kondratieff downgrade, between the peak of price level (at or
near the end of the prosperity phase) and the peak of output (at the end of recession). We
will not, however, insist on this. The idea of trying to establish causal connection in this
way seems to us a mistaken one. No functions connecting two variables only will ever do
in this field. And what covariation there is seems to us much better described and inter-
preted as an imperfect instantaneous one.
Professor Pigou, Industrial Fluctuations, 1st ed., p. 28, makes the unguarded state-
ment that industrial expansions are "invariably" characterized by rising prices and indus-
trial depressions by falling prices. In order to keep this statement from being less true than
the opposite one would be, it must be confined to what Professor Pigou, in fact, calls "nor-
mal" cycles, i.e.y roughly speaking, Kitchins. It could include Juglars only if we replace
"expansions" (i.e., of output) by prosperities. Even so, Professor Persons' criticism in
his review, Quarterly Journal of Economics for 1927-1938, p. 672, is statistically well
founded. But neither seems to perceive the true relation of output to the cyclical phases
in prices, or to realize that no expectation can be formed either a priori or from statistical
regularity as to lag or lead between output and prices. We quote both statement and
criticism because of their methodological interest.
510 BUSINESS CYCLES
Massachusetts figures, though very valuable in themselves,1 start only
in 1889. We shall in this section use English figures only, taking unem-
ployment percentage as an index of the variations in employment. The
figures, which start in 1851, present a sample only — percentage of the
number of members of trade unions reported as out of work2 — and do not
take account of part time. They exclude, however, workmen who were
idle because of strikes or lockouts or illness or age. The series has been
used and compared with others (such as market rate of discount on three-
months bank bills, pig-iron consumption, London clearings, Sauerbeck
prices, increase in bank credits outstanding, rate of real wages, consump-
tion of meat, beer, etc.) by Professor Pigou in Industrial Fluctuations.
In order to save space, we will replace evidence that we ought to present
by a reference to that work, as well as to Professor W. Persons' comments
previously quoted.
We have seen that there is no unique or simple relation between
employment (number of hours worked per week) and output, and that
the latter is not proportional to, or measured by, the former. This is a
consequence of the very nature of economic evolution and becomes
obvious as soon as some of the conditions for proportionality are stated :
production functions would have to be invariant in time and relative
prices of factors would have to be constant. Neither can possibly be
fulfilled for any length of time, such as the period of a Juglar. But both
may be fulfilled approximately in the very short run, for which, moreover,
the second need not always be fulfilled, because adaptation to change in
relative prices of factors may not be possible within it. This very short
run may extend, although it cannot be relied on to do so, to the span of a
Kitchin, which will, therefore, in this respect, present a picture different
from that of the longer cycles. It should also be observed that total
employment or unemployment would in any case be unsatisfactory as an
index of business situations, even if figures were as exact as they are
rough, and even if it were possible to correct them for all the circumstances
that change their significance in time (attitude of workmen to unem-
ployment and to relief, geographical and industrial mobility, length of
working day, and so on). For variation in total employment or unem-
ployment is the net result of what actually goes on in industrial regions
and trades — let alone concerns — and tends to hide differences, which
1 For various estimates of employment and unemployment in this country, including
those of Berridge, Hurlin, Brissenden, and Douglas-Stinebower, see Professor P. H. Doug-
glas's Real Wages in the United States, 1890-1926, 1930, to which the present writer is
much indebted.
2 Professor Bowley made an attempt to go beyond those figures, to which attention
should be drawn. See The Measurement of Employment : an Experiment, Journal of the
Royal Statistical Society for July 1912, in which paper he also shows why, as a measurement
of general employment, our series is anything but trustworthy.
PHYSICAL QUANTITIES. EMPLOYMENT 511
from our standpoint often are what matters most. This may be illus-
trated by the case of seasonal unemployment, which may be considerable
and yet produce but modest seasonal amplitudes, if seasons differ with
different industries.1
We will now inspect the graph of our series (Chart XX) and at the
outset notice the negative covariation of per cent unemployed with the
rate of interest (see also Pigou, Industrial Fluctuations, Chart facing
p. 28), which stands out strongly and is blurred only by the fact that
Kitchins show much more in the graph of the latter.2 Cases such as
the one in 1854, which was obviously due to the dislocations incident to
the outbreak of the Crimean War, do not amount to real exceptions.
For the purpose of analyzing this picture, we will recall or introduce the
following concepts and propositions.
We will call Normal Unemployment the unemployment that would at
any point of time exist if the system had already reached the neighborhood
of equilibrium toward which it is tending. It comprises the cases of
seasonal unemployment which, however, for the reason just mentioned,
need not reveal itself completely and which, as pointed out in the first
chapter, may display a special trend owing to technological change,
such as has in the postwar period occurred in building, or to change in
consumers' habits. It also comprises cases of unemployment due to
such accidents (for example accidental destruction of factories) as will
ordinarily occur, of unemployability (which however we may neglect,3
since, for England and that period, trade unions were not likely to harbor
many such cases) and of unemployment due to change of residence or
occupation or job. Finally, we have seen that imperfections of com-
petition or of equilibrium may account for failure of the system to absorb
all its eligible workmen. This is the only meaning we can attach to the
phrase frequently used in Germany but in a different sense, Structural
Unemployment. With reference to this concept of normal, we define
supernormal and subnormal unemployment. It will be observed that
subnormal unemployment in this sense becomes, outside of the precincts
of perfect competition, compatible with a high figure of unemployment,
1 See Charles Saunders, Importance of Seasonal Variations in Employment in the United
Kingdom, Economic Journal for June 1935.
2 For that reason, correlation between the two series would not be very good. Even if
the relation be judged by the relative number of opposite year-to-year directions of move-
ments, the result is not particularly satisfactory. But this affords only another instance
of the failure of formal methods to work satisfactorily. Everyone will, nevertheless, be
struck by the reality of the relation. In this case a smoothing device would suffice to bring
it out formally.
8 It will be understood, however, that unemployability, extending as it does far beyond
technically pathological cases, is, for other purposes than ours, not only not negligible but
both economically and socially one of the most serious of all the problems of unemployment .
512
BUSINESS CYCLES
PHYSICAL QUANTITIES. EMPLOYMENT 513
that is to say, with a state of things in which many people fail to find work
that they are both qualified and willing to undertake at the ruling rate
of wages. Supernormal employment includes overtime, subnormal
employment short time; but, as stated above, neither can be taken
account of for our period.
A fundamental difficulty must be noticed, although we cannot do
much about it. The various sources that contribute to any given sum
total of unemployment are not independent and their effects cannot be
separated. In particular, the cyclical process affects them all and cyclical
variations of unemployment are affected by them all. This is true also of
all types of normal unemployment and even of unemployability ; for the
question whether a given individual is unemployable or not will be
answered differently in times of good business and in times of bad busi-
ness. Moreover, the same accidental disturbance will give rise to differ-
eiht amounts of unemployment or to unemployment of different duration,
according to the cyclical phase with which it happens to coincide and so on.
Hence, our concept of a normal percentage of unemployment must not
be taken to mean an independent quantity that could be added to equally
independent quantities of other kinds of unemployment, but simply the
percentage of workers unemployed which would exist in the absence of
disturbances of equilibrium. In order to form an idea about its numerical
importance, we must first locate neighborhoods of equilibrium : an average
of observed values would obviously be quite meaningless. 1897 is a
particularly good year to choose since, according to our schema, all our
cycles then passed their neighborhoods. The English figure for that
year was 3.3 per cent, and we are perhaps not very far off the mark if we
consider this value as approximately normal for England at that time.
If, given a state of otherwise perfect equilibrium in a perfectly com-
petitive system, wage rates are raised, say, by public authority, above
their equilibrium value, a definite amount of unemployment will, other
things remaining equal, ensue, which we will call Vicarious Unemploy-
ment. If either equilibrium or competition be imperfect, we may still
use this concept to designate unemployment due to deviation of wage
rates from the figure at which normal employment would be attained,
or — hence our term — unemployment that takes the place of adaptation of
wages to that figure. So far the concept is perfectly definite — though
it may still be impossible to measure it numerically — and simply expresses
a certain aspect of one element of normal unemployment, provided that
state of things lasts throughout the period under consideration, or a
particular type of supernormal unemployment, if it is but temporary.
But we shall also use the term to indicate elements of unemployment
that could be removed by an appropriate change in wage rates, outside
of neighborhoods of equilibrium, And here th§ goncept is, for reasons
514 BUSINESS CYCLES
that will become apparent in our discussion of the behavior of wages,
extremely difficult to handle.
We have above included within normal unemployment such cases as
will currently arise from accidents to firms and are analogous to, say, the
current rate of deaths by accident. If firms suffer economic injury to a
supernormal degree — analogy: deaths from an epidemic — we speak of
Disturbance Unemployment. Many instances of the unemployment
we observe in the course of cycles obviously belong here. Revolutions
or liquidations of wars afford others. From the standpoint of any indi-
vidual industry, a depressive state in another industry also gives rise to
disturbance unemployment. A particularly important class of cases
consists of environments in the process of decline from long-time causes
external to the system. A country being cut off from its hinterland by
the insertion of a new frontier line illustrates what is meant. In this
category we include, when dealing with any single country by itself, the
effects of innovations that develop and run their course outside of it,
for instance, the effects of the innovation of new routes of trade on
Venice after the sixteenth century, or the effects on modern Europe of the
rise of large-scale industry in the tropics.
But for the special case of unemployment arising from disturbance
by innovation within the system we will set up a distinct class, to be
called Technological Unemployment. This term taken literally, of
course, has always been intended to cover only displacement of workmen
by machinery. We make it cover a much wider range and include not
only the effects on employment of every kind of change in industry and
commerce — organizational change, for instance — but also the effects
which changes have on employment in firms or industries that are com-
peted with by the firms of industries that introduce new production
functions. Questionnaires devised to find out from workmen reasons
for their dismissal can, therefore, never bring out the phenomenon we
mean and will always yield results that understate it.1 Yet it should be
clear that our wide definition does but justice to the phenomenon, since
all it does is to extend the usual meaning of the concept to cases funda-
mentally identical in economic meaning and causation to the one com-
monly envisaged under that heading. It is obvious, for instance, that in
the case of replacement of a carriage by a motorcar, the coachman will
1 There is another reason for this. The 1930 Census of Unemployment attempted to
ascertain the causes of dismissal of unemployed workmen and, under the heading of Indus-
trial Policy, inserted the technological cause as one of five subheads. According to the
answers given, the men themselves must have considered all these five subheads of negligi-
ble importance. A little reflection will, however, show that only in a minority of cases
will workmen be able to recognize the technological change responsible for their dismissal.
For this it would be necessary that machines be introduced in an existent plant under the.
eyes of the workmen and that dismissal be effected immediately after.
PHYSICAL QUANTITIES. EMPLOYMENT 515
be technologically unemployed even in the narrow sense, although no
machine drives his horses henceforth, or that it does not make any differ-
ence whether a bookkeeping clerk becomes unemployed because of the
introduction of a calculating machine or another rationalizing device, or
whether a cotton picker becomes unemployed because of the introduc-
tion of a cotton-picking machine or because cotton is being eliminated
by the competition of the standard fiber. Since every kind of unemploy-
ment will induce further unemployment, Secondary Unemployment in
Mr. R. F. Kahn's sense must be added to each.
Now, a large majority of economists will agree in believing that
vicarious unemployment contributes something to the total by which
unemployment varies in the course of cycles (Cyclical Unemployment),
although they do very much disagree about the importance of that
contribution. Our own views about this will be developed in the dis-
cussion on the behavior of wages. All economists will agree to the
statement that disturbance unemployment contributes much. Panics,
spirals, prosperities which facilitate the emergence, and recessions which
facilitate the breakdown, of fraudulent or ill-conceived enterprise, and
all of the phenomena that produce the rise and break of the Secondary
Wave amply account for that. As far as these points go, all we have to
add is that such phenomena would follow from the working of the
cyclical mechanism, whatever the behavior of money and credit, although
autodeflation accentuates them and their effect on employment. But
few, if any, economists realize the one major point that the writer wishes
to make. They have a habit of distinguishing between, and contrasting,
cyclical and technological unemployment. But it follows from our model
that, basically, cyclical unemployment is technological unemployment.
For vicarious and disturbance unemployment are, in the main, but
understandable incidents, though quantitatively important in practice,
which we could abstract from without thereby blotting out any essential
contours. Technological unemployment, however, is of the essence of
our process and, linking up as it does with innovation, is cyclical by
nature. We have seen, in fact, in our historical survey, that periods of
prolonged supernormal unemployment coincide with the periods in which
the results of innovations are spreading over the system and in which
reaction to them by the system is dominating the business situation,
as, for instance, in the twenties and in the eighties of the nineteenth
century.
It further follows that, like profits, technological unemployment is
ephemeral. It might nevertheless be ever present, but, as in the case
of profits, every individual source of it in the industrial organism tends
to exhaust itself, while new ones emerge periodically. In the same sense
as profits, moreover, it may be called frictional, since instantaneous
516 BUSINESS CYCLES
adaptation of the system would kill it at birth. The reader need not
fear that the writer harbors any intention of using these statements in
order to minimize the importance of the phenomenon or the sufferings it
inflicts. But it should be noticed that, in consequence, the primary
long-run interest of the working class is in the effects of innovation on the
total real wage bill and not in the incident variation of employment,
which is but an element of the mechanism that produces the changes
of the former and can be separately handled by public policy.
We now proceed to apply this analysis to the task of explaining the
behavior of our series. Unemployment is one of the primary elements of
our process and, though entirely consequential, also, like everything else,
a conductor of effects which exert secondary causal influences of their own.
The series is natural, systematic, and obviously also cyclical. But our
theory does not suggest any reason for expecting any result trend in
unemployment percentage. We have before us one of the two instances,
among the major elements of our statistical picture, of clean cyclical
series, the other being the series of interest rates. The reader should
carefully reconstruct, from the fourth chapter, every step of the reasoning
that leads up to this statement. Absence of trend in percentage of
unemployed workmen would mean, in the case of a population constant in
number and age distribution, that the process of evolution works in such
a way as to absorb all the cyclical unemployment it creates, technological
and other. If net increase in number of population (natural increase
minus emigration plus immigration) fails to produce a rising trend, this
means that the system also absorbs the current increment at the same
terms, i.e., subjecting it to the same unemployment percentage. Absorp-
tion of emigrants from the agricultural sphere belongs to the first and
not to the second class, even if those "emigrants" were proprietors before,
or for other reasons not technically unemployed.
A glance at the chart seems in fact to verify this result. We observe
no obvious trend, and it is clear that any trend that could possibly be
derived by formal methods would be too weak to be significant, particu-
larly if, recalling what has been said about the untrustworthiness of
individual extremes, we do not attach much importance to the peaks
that occur in 1858 and 1879. It is no doubt true that our statistical
data cannot be relied on implicitly. Among other things, trade-union
figures do not reflect juvenile unemployment and the variations of the
delay in getting first jobs. Unemployment among unskilled and unor-
ganized workmen may have behaved in a different way. Some absorption
could have been expected to occur from the mere effect of such saving as
might have occurred irrespective of our process. Moreover, institutional
changes, insertion into the system of additional permanent rigidities,
emergence at any time of imperfections that had been absent before — for
PHYSICAL QUANTITIES. EMPLOYMENT 517
instance, of the oligopolistic type — could have produced a positive
trend, their elimination a negative trend, so that the basis of our theory
and the basis of our statistical finding are not simply congruent. But it
is not likely that such factors should have combined exactly so as to
produce a spurious verification of our thesis, or that fuller information
would substantially affect it beyond showing up a number of particular
deviations due to particular causes.
As we shall see, cyclical unemployment has within the epoch under
survey been absorbed at increasing rates of real wages. It is this fact
that lends importance to both our result and our observation, which would
become trivial if absorption had been brought about by downward
revision of real wages. For the moment, however, we will merely notice
that, exactly as unemployment may be vicarious to a fall in wages, so a
rise in wages may be vicarious to an increase in employment. As in the
former case we can, theoretically at least, express an increase in unem-
ployment in terms of the equivalent fall in wages that would prevent it,
so we can in the latter case express an increase in wages in terms of an
equivalent increase in employment which is only virtual, of course, since
there was no corresponding labor force in existence. Thus reducing the
long-run effects of the cyclical process on wages to effects on virtual
employment, we arrive at a negative trend in percentage of virtual unem-
ployment, while, if absorption had been brought about by a fall in wages,
the trend of virtual unemployment would be positive.
Expectations as to the cyclical behavior of the unemployment per-
centage are too obvious to be restated. If a cycle started from perfect
equilibrium in perfect competition, increase of employment in prosperity
could show only by overtime if at all. As it is, it partly shows by over-
time— hence, imperfectly in our series. The unemployment of deep
depression will be as irregular as everything is in that phase. It will also
be, owing to the preponderance of what we have called disturbance, of a
sufficiently distinct nature to warrant a distinct name : Depression Unem-
ployment. Revival should bring employment up to normal and actually
does so more frequently than observers assume who neglect the long list of
factors that swell the size of normal unemployment. There seems to be
no theoretical reason why variation in total unemployment should be a
particularly early symptom, but in a majority of cases it actually has
been. Covariation (mostly negative) with all other cyclical series is, of
course, to be expected; but we should always keep in mind that this must
not be interpreted to mean either coincidence of variations or any con-
sistent leads or lags. Economic life is not so regular as that.
The formidable unemployment of the depression phase of the second
Kondratieff shows up well, as does the different state of things in the
prosperity of the third. But considering the nature of the phenomenon,
518 BUSINESS CYCLES
we shall not be surprised to find that it is not so easy to draw two lines of
different gradient through the two Kondratieff sections of the material
as it is with other series. The Kitchins show but little. This also is
perfectly natural — observe the contrast with the Kitchin fluctuations in
interest — considering the importance of the technological component,
which cannot act strongly within their span, and perhaps in part explains
why most students who by business cycles primarily mean Kitchins so
persistently divorce cyclical from technological unemployment. For the
same reason, the violent movement of our series in the Juglar rhythm is
what we should expect. For clusters of innovations, or "steps in evolu-
tion," assert themselves most obviously through Juglars, as we may
also verify by reference to the behavior of the pig-iron series. While it is
interesting to note that, as a rule, each of them tends to absorb the
unemployment of its own creation, nothing can, of course, be inferred
from this about the average duration of the period during which the
unemployed individual is actually out of work. This question has been
investigated for the postwar period,1 but it does not concern us here.
With the exception of the twin peaks in the sixties, which are easy
to understand, there is one peak to about every nine years. Unemploy-
ment above 6 per cent occurs in 16 out of 64 years, and centers, except in
the cases of 1862 and 1879, in Juglar depressions. But it is not coter-
minous with them, lasting longer than they did in three cases (at the end
of the sixties, in the eighties, and in the early nineties), while in several
years which clearly belong to depression phases, that figure was not
reached. Except in the eighties, unemployment never went beyond 8 per
cent for more than one yearly figure. This happened in 1858 and 1879
(neglecting again 1862) and was more than made up for in the sub-
sequent year. Location of the latter peak, at the threshold of prosperity,
is particularly abnormal. Although there is no theoretical reason to
expect even as much regularity of amplitude as there is, it may be safely
asserted that in any unemployment percentage beyond 6 per cent spirals
and so on counted for much more than did the direct effect of innovation.
Troughs, except the one in 1872, display a habit of not going beyond, or
just stopping short of, 2 per cent, which, even apart from the historical
fact that they were associated with overtime, would, according to our
guess, already spell supernormal employment in our (not merely the
statistical) sense. The writer knows of no evidence, that in the epoch
1 See D. Weintraub, Displacement of Workers by Increase in Efficiency and Their
Absorption by Industry, 1920-1931, Journal of the American Statistical Association for
December 1932. The reader's attention is called to this important piece of work, which
by a much more elaborate method arrived at results that are believed, if account be taken
of the conditions of the postwar period, substantially to support our conclusions, in spite
of the failure to connect displacement and absorption with the causation of cycles. We
shall return to this in the last two chapters of this book.
PHYSICAL QUANTITIES. EMPLOYMENT 519
under survey vicarious unemployment played any major role. Corre-
lation between unemployment and wages corrected for price level proves
nothing.1
1 For the unemployment problem of postwar times M. Jacques Rueff investigated that
correlation. On this, see Professor Pigou's analysis, Theory of Unemployment, Chap. X.
In Industrial Fluctuations, Part I, Chap. XXI, he presents the case against causal inter-
pretation of such covariations so forcefully as to make it superfluous for us to argue the
CHAPTER X
Prices and Quantities of Individual
Commodities
A. Prices and Quantities of Individual Commodities (Including
Services). — These should always be studied together. Taken by itself,
neither price nor quantity conveys its full message or in fact any that is
definite, and each must be interpreted in the light of its companion. In
general theory and its statistical complement, this has of course always
been recognized. Students of cyclical variations, however, have some-
times attempted to deal with price variations alone or with quantity
variations alone.1 But the true heroes of the play and the true variables
of the problem are the price-quantity pairs or points.2 Their behavior
1 The above is, however, not intended to convey adverse criticism of such invaluable
works as F. C. Mills, Behavior of Prices, 1927; G. Tintner, Prices in the Trade Cycle, 1935;
or A. F. Burns, Production Trends in the United States since 1870, to which, on the con-
trary, the writer wishes to acknowledge heavy obligation, and which he strongly recom-
mends to the attention of the reader. S. S. Kuznets' Secular Movements in Production
and Prices, 1930, is essentially a price-quantity study and so is a book that in some respects
may be considered as a forerunner of much of the later work in the field and keeps, theoreti-
cally and statistically, a very high level, although it cannot be said to coordinate the theory
of its first part with the statistical investigation that forms its second part — Marcel Lenoir,
Etude sur la Formation et le Mouvement des Prix, 1913. For price dispersion, in particu-
lar, compare M. Olivier, Les Nombres Indices de la Variation des Prix, 1927, and O. Lange,
Die Preisdispersion, 1932. Of course, the whole of the huge literature on price analysis,
although as yet almost unconnected with specifically cyclical problems (the greatest of its
shortcomings and the one most urgent to remedy) really ought to be cited here. The
reader finds a useful guide in the survey of a number of the most significant contributions
presented by L. O. Bercaw, Price Analysis, in Econometrica for October 1934. The
contributions of Henry Schultz, M. Ezekiel, and W. W. Leontief should be particularly
mentioned. See also E. J. Working, Demand Studies in Times of Rapid Economic Change,
and H. Working, Differential Price Behavior as a Subject for Price Analysis, Econometrica
for October 1935. There is also a mimeographed bibliography compiled by the Depart-
ment of Agriculture group. C. F. Roos, Dynamic Economics, 1934, breaks new ground
in various directions. Warren and Pearsons' book on Prices is also relevant to the study
of individual prices. See also the Cornell Agricultural Experiment Station Memoir 142.
This note is of course but a fragment, but it is believed that it gives enough references for
a start.
We shall not insist again on the factual difficulties (as distinguished from difficulties
of interpretation) inherent in the material, which become truly formidable in the case of
even the simplest finished commodity.
2 Such a point (x, y] can be represented by a complex variable, x + yi, or as a vector
520
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 521
is the very heart of the cyclical mechanism. Again a vast research
program opens before us into which we cannot enter. We must confine
ourselves to offering the following comments, which are mere applications
and exemplifications of the analysis of our second and fourth chapters.
1. We will begin by recalling what has been said on the subject of
Group Prices. Most of it applies also to the behavior of price-quantity
relations of individual commodities. Only it does so with much greater
force because individual peculiarities must relatively increase in impor-
tance as we proceed from textiles to cottons, from cottons to cotton
wearing apparel, from this to cotton shirts, from cotton shirts to men's
cotton shirts, and from these to brand X of firm F. Again, we may con-
veniently, though somewhat artificially, class variations into those which
are directly caused by innovation (and these in turn into the price-
quantity variations of the commodities in the production of which
innovation has occurred, innovating commodities as we call them, and
the price-quantity variations, induced by innovation, in their rivals and
subsidiaries) and those which are simply responses to the general business
situations that result from the particular disequilibria created by
innovation. It is with the latter class that such statistical uniformity as
there is must rest. And in fact the impress of cyclical phases is in the
large majority of cases clear enough, although not so strong as students
expect who have acquired, from an aggregative view of the cyclical
process, a habit of mind for which any deviation is in the nature of a
discovery to be wondered at. But every industry or even concern has a
structural pattern of its own and is at every point of time faced by an
individual! set of environmental conditions that determines its reaction.
Hence if we observe that the price-quantities of an industry behave
differently from those of other industries, this is no reason for concluding
that different forces must be acting upon them or that they have (in any
but a formal sense) a special cycle of their own, or for denying the reality
of the " general" cycle.
There is much danger of misinterpreting statistical findings in that
sense, especially if we refuse to recognize the traces of the cyclical process
unless they display sufficient regularity to reveal themselves to formal
methods of measurement. The reader should therefore realize from the
outset that there is no warrant for this attitude and that expectation from
our model is not for uniformity but for what we actually find, great
variety1 of amplitudes, periods, and sequences that does not tell in the
since complex variables may be geometrically interpreted as vectors from the origin. Pro-
fessor W. W. Leontief's important paper on Price-Quantity Fluctuations in the Business
Cycle (Review of Economic Statistics for May 1935) seems to the writer to open up a most
promising as well as novel line of approach.
1 "Dispersion graphs" on which lines representing prices or quantities on a time scale
are made to coincide for the first year are useful to inspect in order to get a first impression
522 BUSINESS CYCLES
least against the presence of an all-pervading movement and does not spell
theoretical, although it does spell statistical, irregularity. He should
train himself to look upon every industry as a resonator exposed to the
impact of a force and responding to this impact according to its structure.
These resonators are differently affected by each individual impact,
although it is always the same process that impinges upon them, and they
would, because of this, respond differently, even if they were equally
constructed. But they are not and hence would respond differently
even if equally affected. It follows that lead or lag of individual price-
quantity pairs as compared with price level or total output, as well as
compared with other price-quantity pairs, has but little significance for
the causation and mechanism of the cyclical process per se.
£. We may still schematize, however. And first we will attend to the
behavior of the price-quantity pairs of industries which, in the sense of
the above distinction, are not innovating ones or close relatives to these.
That distinction is not easy to carry out in any actual instance because,
as we know, innovation jumps from industry to industry, and intrudes
into every one of them at some time or other. But it has some expository
virtues. It is convenient, moreover, to start with cases that approximate
the competitive schema.
The impact of the cyclical sequence of business situations on such
industries is by way of enterpreneurs* expenditure and expenditure
induced by it. Effects may be decomposed into nominal and real ones.
By the first we mean effects on prices and costs that cancel each other
and by the second, effects that do not. Although it would mean eliminat-
ing the very essence of our process if we adopted the hypothesis that that
expenditure so acts as to affect all values in equal proportion,1 part of
what actually happens does answer to the pattern of such a purely nomi-
nal alternation of "inflation" and *' deflation,*' which should induce
neither expansion nor shrinkage of output, and in fact explains why varia-
tions of prices have, in all phases of cycles, much more in common than
have variations of quantities. There is some point, therefore, in correct-
ing commodity prices by an index of the level, though the defects of these
of that variety. See, for example, the charts on pp. 12 and 18 of Professor Irving Fisher's
The Making of Index Numbers. The picture is, in its pedagogical value for general pur-
poses, somewhat injured by the fact that it covers the years of the World War. Quantities
are seen to be even more individualistic than prices. For a picture of average rates of
change in prices (Professor Mills calls them "trends") see F. C. Mills, Behavior of Prices,
p. 68. The graph has the virtue of starting not too far from a neighborhood of equilibrium
and of illustrating a Kondratieff prosperity. What might be termed a modal behavior is
discernible.
1 In that sense it is of course true that there is no "one-to-one relation" between any
single price and the price level, and that in any multiple correlation analysis price level
should, hence, be explicitly introduced as one of the independent variables.
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 523
indices make it a hazardous operation. This has been done by Mr. Carl
Snyder (Structure and Inertia of Prices, quoted before, p. 192 for the
1840 45 1850 55 I860 65 1870 75 1880 85 1890 95 1900 05 1910 13
CHART XXI. — United States individual prices deflated (see Appendix, p. 1058).
United States, and p. 194 for Great Britain). He, however, seems to
claim for the results an absence of long-time tendency and a covariation,
which not everyone will be able to see.1 Chart XXI will convey an
1 However, a tendency to covariation may be said to exist also in the deflated figures.
But our version, both of this finding and of the covaration of undeflated figures, may be
given in the form of an answer to Mr. Snyder's question as to what mainly determines
prices: the cycle.
524 BUSINESS CYCLES
impression that suffices for our purpose. With some obvious qualifica-
tions, variations in the sum total of money wages would, under our present
hypothesis, reflect those variations, both in costs and receipts, which
would cancel out from the standpoint of the firms and the wage earners
themselves. In reality they cancel out to some extent and, so far as
they do, we have no difficulty in taking account of the cyclical element in
an ordinary Marshallian demand or supply curve by introducing, say,
the wage bill as a parameter that will make it shift in the requisite way.1
From the standpoint of some theories of the cycle, this is in fact all that
would be needed.
3. But as we know, increase and decrease of expenditure do not act
in the way just envisaged but so as to produce shifts of "purchasing
power" between households and shifts of relative receipts and costs
between industries and even firms. There are, therefore, real changes
to which to respond and this response is, with but minor qualifications,
uniquely determined in the competitive case. Only, deterrninateness
does not mean that every industry and firm will respond in the same
manner. It will do so — and the analysis of the second chapter prepared
us for this — according to its structure and its situation at the moment.
By structure we do not only mean the technological and commercial
setup — including the lags inherent to it — and the organization of the
industry — such as its relation to, and the behavior of, its wholesale and
retail trade2 — but also the nature of its products — for example, whether
they are subject to rapid change in fashion, readily variable, easily stor-
able, or not — its financial habits, the support it can expect from its finan-
cial connections, finally, the mentality, the promptness, and particularly
the horizon, of its managerial personnel, its aversion to dismissing work-
men, and the like. A most important point is the policy of its firms
with regard to stocks, for, given the possibility, producers will in general
"speculate" in their own product. Data about carry-over of stocks
being exceedingly scarce for prewar times (although valuable indications
might be unearthed), we have no choice but to neglect this element of
cyclical price-quantity fluctuations.3
1 If we take a linear demand function (entirely inadmissible and even absurd though
it is) that expresses quantity (virtually) sold, y, in function of the price x and the "catch-
all" time, y — a — bx + ct, we may either interpret the last item to mean cyclical varia-
tion of the kind discussed above and replace it by cw, w meaning wage bill, or we may leave
ct for other "trends" and let a vary in function of w, in this case simply as Kw, K being a
constant. Of course we can in this case also correct for price level and then use the original
equation,
2 The very different attitudes of retailers toward their margins may be cited as an
example. As a rule, these margins are not rigid, but in some branches they are.
8 It is, however, not too hazardous to draw conclusions from postwar material. The
study of R. H. Blodgett, Cyclical Fluctuations in Commodity Stocks, 1935, and various
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 525
Behavior of price-quantity pairs also depends, at any moment and
during every individual cycle, on the actual situation of each industry
and firm at that time and in principle on all preceding situations. What
we mean by situation would be very easy to define if we could start from
perfect equilibrium. But this is precisely what we cannot do now.
We must start from the second approximation of the fourth chapter and
take into account all the underemployment, all the sloppiness, all the
effects of indebtedness, all the undigested leavings of previous phases,
all the imprints of chance events peculiar to the industry under study,
that may be present and account for the diversity and the timing of its
response. Cycles may then be skipped or there may be what to the
statistician will look like special cycles, although the industry may be
reacting to nothing but "the" business cycle itself. The moral of the
story is that only analysis of the history and state of an industry will
explain the behavior of its price-quantity pairs. Unless research into
the facts can proceed on these lines it is useless, and we might just as well
confine ourselves to stating our theoretical expectations, for the statistical
graph will not tell us more than they.
4. However, these are only qualifications, although important ones,
of the broad truth that noninnovating commodities which arc produced
either under conditions of perfect competition or under conditions which,
even in the absence of technically perfect competition, enforce similar
results will display cycles in prices both relatively promptly and rela-
tively strongly and hence, on the whole, tend to fluctuate less in quan-
tity, unless typically cyclical in nature. In order to verify this we need
only glance at the constituents of some of the indices of sensitive prices.
The Snider-Persons index of %Z commodities (1875-1889), the Harvard
Economic Society index of 13 (from 1923), and another Persons index
of 10 commodities (1890-1922) — for all of them see W. M. Persons,
Forecasting Business Cycles — list beans, barley, burlap, coffee, cotton-
seed oil (refined), cotton sheetings, coke, copper (ingots), corn, corn
meal, hides, hogs, lard, oats, pig and bar iron, pork, print cloth, rubber, rye,
shellac, silk, spelter, steel scrap, tallow, tin, tin plates, tobacco (leaf), tur-
investigations of the Berlin Institut filr Konjunkturforschung (for example, the article
on Lagerhaltung in the Vierteljahrshefte, 1928) are particularly suggestive. To the late
Professor Laurits V. Birck of Copenhagen is due the idea, the present writer believes, of
making it a criterion of cyclical situations where stocks are primarily held — with producers,
wholesalers, retailers, or ultimate buyers. The pure theory of the subject is largely the
work of Professor Tinbergen. Mr. Blodgett's results amply bear out the expectation
that, owing to the variety of individual situations and of relevant considerations, behavior
as to the holding of stock must differ widely between industries and firms. For instance,
stocks in goods that are perishable or subject to fashion will of course be positively
associated with cycles, stocks which are due to inability to sell and to the difficulty or
costliness of varying the rate of production, negatively.
526 BUSINESS CYCLES
pentine, wheat, wood screws, wool, worsted yarns, zinc. Hemp adorns a
corresponding German list. As the reader sees, the expectation of finding
those commodities, which according to some theories are particularly
allied to cyclical processes, is not entirely disappointed. But agricultural
articles, the standard examples of almost perfectly competitive condi-
tions, predominate. They would do so still more were it not for two
circumstances. First, though prices and quantities of all of them are
of course strongly influenced by chance variations of crops, some are
more than others. Potatoes, for instance, are dominated by this factor
much more than wheat (see H. L. Moore's highly satisfactory demand
curve for potatoes in Economic Cycles, p. 76, which fits well, in spite of
the extreme simplicity of the assumptions underlying its construction —
the price of wheat, for instance, shows much more the influence also of
other factors) . Second, the purpose of deriving a sensitive index which
precedes the index of the general price level, disqualifies others the prices
of which, like that of milk, display the influence of the cycle all right.1
1 Professor Mills accuses barley, beans, corn, rye, wheat, and many others of irregular-
ity. This is, of course, a correct formulation of the results of applying his measurements
and his reference schema of cyclical movements. Others (26 in all) he classes as "excep-
tional," excluding them altogether. But his warnings, op. cit., p. 102, as to the interpreta-
tion of Table XV in his Appendix (and others) are but too justified. It might be added
that those results afford a good illustration of our view about such measurements. If
irregularity of behavior be interpreted to mean (for example, p. 553n.) that the prices of
commodities so designated "do not conform in any systematic fashion to the cyclical move-
ment of general prices" and, we suppose, "exceptional" behavior to mean irregularity too
great to be borne with, it seems tempting to infer that the price-quantity behavior of such
commodities is independent of the cycle or contrary to the cyclical schema. This, of course,
Professor Mills does not mean, and it would be completely false. That some of them could
ever have been drafted for service in an index of sensitive prices and that they served better
than pig iron, is in itself sufficient to refute it. Another argument is afforded by eggs and
butter, also classed as irregular, and poultry, which is voted exceptional. Eggs, butter,
poultry, as well as all the meats, classed as regular, beef, pork, mutton (the writer is not
sure about veal) display in their price-quantity behavior as close a relation to wage bill as
we can expect, taking account of the structure of these "resonators." Again we must for
this statement mainly rely on postwar data. The wholesale price of butter moves inversely
to Federal Reserve Board Pay Rolls in 1927, but otherwise covariation prevails. The gross
Farm Value of Poultry and Eggs, as estimated by the Department of Agriculture, goes well
with Pay Rolls. Action of variation in consumers' expenditure through wholesale prices
on farm prices is comparatively quick in all cases. See, for example, United States Depart-
ment of Agriculture, report on hog situation to United States Senate, Feb. 9, 1933, which,
though dealing with postwar fact, may still have application to our period: "total con-
sumers' expenditure for pork apparently is determined largely by the level of consumers'
incomes. Consumers' expenditures for a large supply of pork are about the same as for a
small supply if consumers' incomes remain constant." The same is approximately true
of the other articles mentioned. To be sure, this will result in a variability of price which
may look irregular in the sense alluded to, while, as a matter of fact, it would be difficult
to find much stronger instances of response to cyclical impulses. Similar remarks apply to
some of Mr. Tintner's measurements. He states, for instance, op. cit., that German metal
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 527
The nonagricultural articles call, in this connection, for only two com-
ments: cotton sheeting, print cloth, and worsted yarns are not in the
least less sensitive than pig iron or copper. And wood screws are par-
ticularly interesting as an article the Marshallian elasticity of demand
for which must be extremely small and the production of which is carried
on under conditions certainly far removed from perfect competition.
Professor Mills gives 40.2 months for the average duration of its Kitchin
(op. cit.y p. 545, No. 291; bar iron, No. 267, has 39.1).
This is not to deny that a more intensive study of periods, amplitudes,
and relative phases (timing) may yield results that reveal cyclical causa-
tions. It is not uninteresting to note, for instance, that metal and metal-
product prices and quantities indicate Juglars much more clearly than
Kitchins (so, however, do pepper and flour, also freight rates1 and some
chemicals; see Tintner, op. cit.), and that foodstuffs and raw and semi-
finished textiles and textile machinery do the reverse (as, however, also
do other chemicals and timber). But no simple generalizations are
possible and no formal methods of analysis are satisfactory. As stated
above, this entirely conforms to expectation from our model. For the
study of demand and supply it means that Marshallian curves fail
beyond remedy. In specially favorable cases we may still split the
actual movement into a movement, say, along a demand curve, and a
movement of the demand curve. The potato case cited affords an
instance in which the first so dominates that, for quiet times, we can
derive what looks like a plausible Marshallian curve. Even there it
does more to obscure than to elucidate the real course of events. The
cases of iron, steel, copper, and other commodities of that class afford
instances in which the second so dominates that we get II. L. Moore's
positively inclined demand curves, which, as is universally recognized
now and as was eventually recognized by Moore himself, are no demand
curves at all, but paths of cyclical shift within a family of demand curves,
each of which is still negative to the quantity axis. But to reason as if
cyclical variations of quantities supplied or demanded were movements
along any invariant curves is obviously inadmissible and, as will be seen
in later chapters, at the bottom of many faulty arguments about wages,
interest rates, and prices of capital goods. It is no less inadmissible to
assume that, while shifting, the forms of those Marshallian functions
prices show small response to the cyclical rhythm. His method yields this result. But the
reader should look for comment to the chart on p. 96 in the price study, quoted before, of
the Berlin Institut ftir Konjunkturforschung. Such procedure makes it still more difficult
to segregate the cases that really present difficulties of interpretation, such as tea, which the
present writer entirely fails to understand (see Mrs. Gilboy's articles on Time Series and
the Derivation of Demand and Supply Curves in the Quarterly Journal of Economics for
August 1934).
1 (English) prices of secondhand ships move particularly well in the Juglar.
528 BUSINESS CYCLES
are invariant to cyclical phases. This is much more important than are
many of other objections usually raised against the use of "classical"
analysis in this field.
B. Special Cases. — Recent work has gone beyond those limits.
Rate of change of price,1 the influence of past prices,2 and a number of
collateral circumstances have been taken into account (Evans, Roos,
and others. See for a survey of formulae sufficient for our purpose,
and for a highly instructive application to steel, R. HH Whitman, The
Statistical Law of Demand for a Producers' Good, Econometrica for
April 1936). The problem of the influence of variations in real income
has been analytically taken care of by the Pareto-Slutsky-Schultz
theory (see the latter's paper in the Journal of Political Economy for
August 1934). Multiple correlation has lent its aid for advance in
similar directions (Ezekiel, Bean). Reaction of consumers to change
in price and income has also been the subject of a League of Nations
investigation (see on part of that project, Mr. H. Staehle's article on
the behavior of immigrants into the United States, Econometrica for
January 1934). Some of the results of all those and similar endeavors
directly fulfill our desiderata. Others are readily adaptable to the
purposes of the study of cyclical behavior and in fact may be improved
by borrowing from the theory of cycles. All break down when produc-
tion and consumption functions ("methods" and "tastes") change,
as they unavoidably do eventually. But for most commodities they,
particularly the latter, remain constant for considerable stretches of
time. Any change that occurs by way of innovation or induced or
autonomous change in taste must be located historically. In some cases
it shows in the statistical material, which, however, must never be
relied on to indicate explanation unequivocally. Had we nothing but a
historical graph of quantity of products from the soil, we might easily
be tempted to interpret its form in the sense of decreasing returns,
absurd though that would be.
1 Professor Evans was, as far as the writer knows, the first to introduce rate of change of
price explicitly into the demand (and supply) functions. The simplest way of doing it is
to add an item of the form hp, h being a constant and p being the time derivative of price.
See, for example, his Mathematical Introduction to Economics, 1930, and earlier papers
This may be interpreted as a way of taking account of anticipation. We shall use this
idea later on in another connection. That it works with speculation in agricultural com-
modities has been shown by investigations of the Department of Agriculture group.
2 If we treat quantity as a function of, among other things, past prices — say, an average
of the prices of the preceding five or six years — we may mean either that the capacity
adjusted to those prices will persist for a time or that people really intend to react to those
past prices. Formally, this makes no difference; but it follows that, if we observe that
farmers react in a way which may be amenable to interpretation in the latter sense, we
must not from this alone infer that they really mean to do so.
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 529
We may take another step without leaving the precincts of our
present hypothesis of (more or less) perfect competition, although the
element of innovation will occasionally intrude. Of the reasons why
fluctuations of price-quantity pairs of a commodity may fall out of line,
in a manner not explicable by the cyclical mechanism without taking
account of the facts of industrial structure, technological lag is one of the
most important. To some extent it is ubiquitous, particularly if we
include adaptation by construction or reconstruction of plant. But we
will confine ourselves to a few outstanding instances. Coffee is one.1
We shall expect and actually find in the series of its price-quantity pairs
traces of the time shape of the yield of coffee trees. The American price
— we deal with Brazilian coffee and the United States market only — varies
fairly closely with the Kitchin phases (average duration for 1890 to 1925,
according to Professor Mills, No. 109, 37.0 months), but all the longer
movements are entirely out of step moving under the obvious influence of
enormous waves of new planting. New trees begin to bear after from
four to seven years, after which their yield varies with weather conditions
and the exhaustion of trees after bumper crops. Such a wave pulled
the price down to a trough in the middle eighties and again, after spec-
tacular rise to the early nineties, at the beginning of this century.
Interpretation is not difficult but complex. It was a case of innova-
tion, and of an article forcing its way into consumers' budgets and
inducing a change in taste : per capita consumption increased systemati-
cally to about 1902, obviously not only in consequence of increasing
wealth, as we may infer from a comparison of the changes that occurred
in the relation between coffee and tea consumption in this country and
England. Production being mainly for export, international influences
caused a deviation from standard patterns. But the process was further
distorted by monetary disorders in Brazil, which frequently put an
additional premium on production, and by the policy of government,
which subsidized it in various other ways even before the Difesa do Cafe
set in. There is also the influence of weather, of the development
(primarily, 1870 to 1890) of steamship lines, domestic transportation
(until the sixties, coffee was transported from plantations to ports by
mules), the development in the importing countries of subsidiary indus-
tries, and the changing competitive situation with relation to the other
producing countries and to tea. If we take deviations of price corrected
for level and of per capita consumption from curves drawn freehand
1 Of the considerable literature we will mention only H. Roth, Die Uebererzeugung in
der Welthandelsware Kaffee im Zeitraum von 1790-1929, and E. W. Gilboy's Study of
Coffee and Tea, 1850-1930, previously cited. Also see Rowe, Studies in the Artificial
Control of Raw Material Supplies, London and Cambridge Economic Service, Special
Memorandum 35, 1982.
530 BUSINESS CYCLES
through points of inflection of smoothed original data, we find good
inverse covariation (see Mrs. Gilboy's chart, op. cit.9 673), which shows
that the "movement along a demand curve " was not entirely extinguished
by shifts. But the line through the inflection points of the price graph
has a shape all its own. The corresponding line for per capita consump-
tion since about 1870 displays only what may be called a result trend.
The example displays very well the list of problems we encounter in
price analysis and particularly, the writer is grieved to say, the difficulties
which exact methods of analysis are sure to meet. Technological lag,
by displacing and distorting effects and giving large scope to miscalcula-
tion, certainly will produce fluctuations which would be absent without it
and cycles that are special in this sense. But first, it would be a grave
error to assume that the fluctuations actually observed simply indicate
the effects of the presence of technological lag in adaptation and, second,
it would be not less erroneous to think that we have here the case of an
endogenous fluctuation, which of itself might go on indefinitely, possibly
even with increasing amplitude. Coffee simply responds to a great
number of impulses. It creates cycles if, as, and when its production
involves innovation; it experiences the influence of cycles as far as it
experiences cyclical variations in consumers' expenditure. The latter
effect is, if anything, more clearly marked than the effect of the coffee-
tree lag. And it is to these cyclical and external influences that its
fluctuations are due. Only the form of these fluctuations is shaped by
the structural properties of the coffee resonator, of which the lag is
one.
The point we are trying to make stands out still more clearly in the
"cycles in animals." The case is somewhat simpler because, though
there are plenty of qualifications, production of animals for human
consumption is, for Germany and the United States, primarily a domestic
industry and, again with exceptions, one in which the element of innova-
tion does not so powerfully disturb the picture as it did in the case of
coffee. These cycles are also held to derive from the fact that adaptation
to any supernormally or subnormally favorable situation is possible
only with a technological lag, corresponding to the time it takes to rear
an animal to the requisite age, substantially fixed1 and the same for all
1 As the reader sees, we are simplifying to the utmost. None of the above statements
is entirely true. To begin with, even if the period of rearing and fattening were exactly
equal for all the farmers of a country, the moment when they get ready with their supply
would differ somewhat because in some cases and within certain limits the decision, for
instance, to rear more hogs can be immediately given effect to, while, in other cases and
beyond certain limits, time-consuming rearrangements on the farm may be necessary.
Second, the situations facing farmers, particularly in different parts of a country, are not
equal. The effect of a favorable or an unfavorable fodder-meat ratio depends on the
possibilities in other lines of production. Third, there is, both for the initial decision and
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 531
producers who, being faced by the same situation, have to make their
decisions at about the same time. It should be observed at once that
it is further necessary to assume that they will take no account of their
competitors' actions until they all of them come out with their product.
Obviously, this assumption is unjustified, for the behavior of competitors
is no secret. Therefore, waves started by a single disturbance would,
notwithstanding the lag, soon die down. The idea that one disturbance,
once having caused an exceptionally favorable or unfavorable situa-
tion, will thereby create a wave that might, under its own steam as it
were, go on forever is manifestly absurd. Producers may and actually
do react in those cases in such a way that the mass effect of their adaptive
action will create a disequilibrium in the direction opposite to the devia-
tion to which they intended to adapt themselves; and rebound from the
untenable position is similarly likely to outrun the goal. But eventually
they would learn the lesson. It is nothing short of thoughtless to trust
the constants of some equation to supply proof of the contrary. We
recall what has been said on the subject in the second and fourth chapters
and conclude that, since those waves go on apparently indefinitely and
since they are (comparatively) so regular, they can precisely not be
endogenous in the full sense of the word, but must be kept going by
shocks from without or by a generating mechanism. This mechanism
can be only the cycle — particularly, the Kitchin, but also the Juglar —
acting through consumers' expenditure.
That this is so can be readily seen, for instance, in the standard case,
the hog cycle.1 Hog prices, on the whole, move well with all three cycles
later on, room for choice and also for revision of choices made. Even if all had actually
produced the same type of animal with a view to having it ready at about the same date,
necessity for selling at that date differs widely. But although these and similar considera-
tions should make us less confident in the interpretation of certain striking regularities,
we may take it that, for the limited purposes we now have in view, they are of secondary
importance
1 But it is the same with cattle and lambs. All those cases are complicated by inter-
commodity relations and jointness in animal products and, of course, such external events
as weather (including its effects on the price of fodder), variations in consumers' tastes, and
so on. Sheep production is nevertheless, in this country at least, a very good case, in spite
of the complicated relation of the consumption of lamb and mutton to that of beef, pork,
poultry, and veal (to the last of which it shows some inverse relation) . The sheep popula-
tion in this country did not display any marked tendency from 1890 to 1913, but three very
regular cycles. Recovering from a drop it rose to a maximum in 1893, dropped to a
minimum in 1897, reached an almost equal maximum in 1902, and a somewhat higher one
in 1909, after which it dropped sharply. Farm value of sheep per head has maxima in
1893, in or about 1900 and 1906, Chicago price (medium to choice) in 1892, 1898, and 1905.
Sheep price precedes sheep population by about two to three years. There was, of course,
the transition from production for wool to production for meat and the increasing emphasis
on lamb production. Those six-, eight-, or nine-year cycles cannot be simply linked to the
lag; on the contrary, they point to other factors. They seem to have much more to do with
532 BUSINESS CYCLES
(see, for example for Germany, the chart on p. 91 of the new price study,
previously quoted, of the Berlin Institute, covering by two scries pork
till 1863, hogs since — nearly 150 years. For this country and over his
period, Professor Mills gives a good mark to his two hog entries, Nos.
15 and 16, duration of cycle, 38.8 and 38.4 months, respectively). Their
relation to cyclical changes in consumers' income is too obvious to
require proof.1 It is true that these shifts of demand are combined
with strong movements "along demand curves." Receipts of hogs at
markets or, for that matter, slaughter, are related inversely to prices of
hogs (see, for example, Mr. Hanau's graph for Prussia, 1900-1913, p. 18
of the first edition of his study) and positively to the relation between the
prices of hogs (or pork) and of fodder that prevailed about 18 months2
earlier. Prices of hogs could almost be forecast from this relation alone,
and it therefore seems at first sight that that market is entirely dominated
the sweep of the Juglar, although the relation is as blurred as we might expect it to be.
Mr. M. Ezekiel's most interesting study, Factors Related to Lamb Prices, Journal of Politi-
cal Economy for April 1927, entirely free from special-cycle obsessions, and a landmark for
that kind of work, finds a 68.3 per cent " determination" of the price of dressed lamb (in
the multiple correlation sense) by Mr. Snyder's General Price Level. This is quite enough
to establish our point, although correlation with a pay-roll index would have been more rele-
vant. That business activity figures among determining factors only with 0.7 per cent
does not prove anything against us, because it is measured by the Harvard Price Index of
Business Cycles, deflated by the B. L. S. wholesale price index. Mr. Ezekiel himself
expresses doubts on the subject in his note, op. cit., p. 248.
We cannot and, for our purpose, need not go into the ever-expanding work on hogs.
We merely quote Haas and Ezekiel, Factors affecting the Price of Hogs, Department of
Agriculture Bulletin 1440, November 1926; (S. Benner, the discoverer of the hog cycle,
as far as the writer knows, should not be forgotten however: Prophecies of Future Ups and
Downs in Prices, 1876); Sarle, Forecasting the Price of Hogs, American Economic Review,
1925; Sewell Wright, Corn and Hog Correlations, Department of Agriculture Bulletin 1300;
A. E. Taylor, Corn and Hog Surplus of the Corn Belt, 1932; and A. Hanau, Die Prognose
der Schweinepreise, Vierteljahrshefte zur Konjunkturforschung, Sonderhcft 2, revised ibid.
18. For cattle, see the same author's study, ibid., Sonderheft 13. Grateful acknowledg-
ment should be made for helpful suggestions derived in this whole range of subject from
L. H. Bean, The Farmer's Response to Price, Journal of Farm Economics, 1929.
1 See O. V. Wells, Farmers' Response to Price in Hog Production and Marketing, Depart-
ment of Agriculture Technical Bulletin 359, April, 1933, in which study this element has been
considered, p. 8. Notice also the behavior of commercial slaughter, p. 37. Haas and
Ezekiel, op. cit., arrive again at a very low estimate of the influence of the cycle by taking,
this time, not only the Harvard Price Index, but also an index of prices of industrial stocks
as a series to correlate with, for which there is still less justification. The graph on p. 26
of that same study shows the Kitchin almost ideally.
2 There was some "technological change" in fattening which seems to have affected the
period, but it was about that length in the last prewar decade both in Germany and in the
United States. It is, however, not quite convincing, for the period of gestation plus the
period of rearing and fattening is more nearly 15 months. That the farmer does not act
at once upon a given hog-fodder price relation is plausible. It is less plausible that it should
always take him just 3 months to arrive at a decision.
PRICES AND QUANTITIES OF INDIVIDUAL, COMMODITIES 533
by the variations of supply, which in turn are entirely mechanical. The
inference about the presence of a special cycle unrelated to the business
cycle which some students draw is, nevertheless, entirely unwarranted.
What we behold when looking at those hog graphs that are so remarkably
regular, is nothing but the — wavelike, to be sure — working of a particular
apparatus of response.
C. The Cycle in Shipbuilding. — This cycle, made famous by Professor
Tinbergen,1 serves to illustrate a lag phenomenon incident to all time-
consuming construction of plant and equipment and therefore differs
(also in other respects) materially from the hog case. Our discussion
continues an argument opened in the fourth chapter, sec. E. Let us
consider total tonnage as a certain function of time (cyclical time in the
first instance, but later to be linked with historical time), say, /(£); and
let us identify rate of change in total tonnage in a first approximation
with shipbuilding, which therefore is f'(t). If, at any point of time, total
tonnage be above normal — whatever that may mean — this will send down
freight rates and reduce shipbuilding (it cannot make the increment of
tonnage negative). If total tonnage be below normal, the reverse will
happen and after a period #, determined by the time it takes for shipping
companies to order and for shipbuilders to build new tonnage, total
tonnage will increase according to the intensity with which carriers react.
This intensity is assumed to be a constant a, and is measured in terms of
the increase or decrease in tonnage ordered that corresponds to some unit
deviation of tonnage from normal. Shipbuilding or rate of change in
tonnage is thus linked to that tonnage which existed at time t — $,
hence f'(t) = ~af(t — $). This functional equation is treated in the
familiar way, i.e., by substituting.
f(t) = eat+P = Ceat
Clearing for Ceat, we get a = — ae~a*> and if we put — <*$ = x + iy> we
get an exponential with a complex exponent the imaginary part of which
will give us periodic fluctuations.2 These fluctuations in tonnage can
1 Ein Schiffbauzyklus? Wdhvirtschaftliches Archiv for July 1931. What will be said in
the text should not be construed as adverse criticism of Tinbergen's work. In later papers
he has shown, in particular, that he is fully aware of those elements of the case that we are
going to stress.
2 There is some danger of the layman's misunderstanding this procedure. The complex
exponent is, of course, chosen because we know that there are fluctuations in the phenom-
enon. But that these fluctuations are due to the lag that enters into the functional, from
which they are then by means of that complex exponent deduced, is really a new hypothesis,
which, to be sure, may derive justification from good fit of results — we do not believe it
does — but which must not be mistaken for a result following from the original setup itself.
This could equally well be satisfied with an aperiodic solution, which the existence of the
fluctuations would not prove to be wrong. They would simply have to be ascribed to
something else.
534 BUSINESS CYCLES
then be represented by a composite of cyclical and of aperiodic move-
ments, periods depending on & and a only. The composite, cleared of
meaningless solutions, is extremely pliable, the introduction of this piece of
apparatus a matter of congratulation.
But, as applied to our subject, this chain of reasoning gives rise to
various doubts. To begin with, fluctuations of this kind cannot be called
endogenous in the sense of being self-generating, for they obviously
depend for their existence on some disturbance that starts them. Second,
they can be self-perpetuating only if we assume a very peculiar form of
reaction by carriers, which violates not only the general assumptions of
economic theory about rationality of behavior but is at variance also
with commonsense. We cannot reasonably assume that reaction to,
say, abnormally favorable freight rates will be mechanical and proceed
without any consideration of the causes and probable duration of that
state of things and of the effects of simultaneous action by the whole
trade. It is no valid objection that results will only show with a lag and
impinge upon a situation that has changed meantime, so that even
reaction correct ex visu of the moment in which it was decided on may
yet be proved by the outcome to be wrong : for this would not happen if
no other process were at work or if no further external disturbance
occurred — in which cases it is that process or this disturbance and not
any automatism inherent to shipbuilding that causes the discrepancy and
further fluctuations. Third, the argument involves neglect of the
regulating influence of a reaction of prices of new ready ships, second-hand
ships, and newly ordered ships. Fourth, there is the objection to using a
trend line as a normal from which to measure deviations of tonnage.
Since this trend itself is the work of growth and cycles, its elimination
removes part of the essence of the process. Nor can it, fifth, be taken as
a standard by which to judge correctness of reaction, for cyclical varia-
tions of demand must be taken into account both by carriers and by
the economist who observes their behavior: freight rates are not an
invariant function of tonnage alone and, as soon as this is recognized,
there is an end of this particular cycle.
Inspecting Chart XXII, we find the implications of this fully borne
out by the facts of the case. Of course we see the effects of all the
peculiarities of the industry — the effects of subsidies and policies of
prestige embarked upon by other nations but influencing the British
situation; of the fact that shipping and ship-building reflect economic
and political conditions all over the world; of the relation that trans-
atlantic freight carrying bears to emigration; of the fact that shipping will
share, to a large extent, the fate of international commodity trade, which
itself deviates, especially for longer periods, from the contour of the
cycle; and so on. Above everything, we must attend to the fact that,
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 535
technologically and commercially, important innovations took place in
shipping and shipbuilding, which will account for the long-time tendency
of freight rates to fall even after the general rise in prices had set in, and
partly also for rate wars. Considering all this, cycles show well enough
both in rates and, particularly, in construction. Their movements,
ANNUAL :.
,' INCREASE:
: IN TOTAL:
TONNAGE ;
PRICE "OF NEW,
READY CARGO
1870 1875 1880 1885 1890 1895 1900 1905 1910 1915 1920 1925 1930 1934
CHART XXII. — British shipping and shipbuilding (see Appendix, p. 1059).
inverse to each other in a long-time sense for the period covered by the
chart, and substantially synchronous and in the same direction in a
short-time sense, are exactly what we should expect in the case of the
cyclical process acting on this particular resonator and can be fully
accounted for without the element of lag. The writer believes, in fact,
that the latter had very little influence and that such verification of the
opposite view as may be derived by formal methods is largely delusive.
D. Entrepreneurial Price Policies. — If the entrepreneurial act con-
sists in the production by a new method of a commodity already produced
under conditions of perfect competition which the entrepreneur is powerless
536 BUSINESS CYCLES
to alter, so that he is confronted by an individual demand curve of
infinite elasticity, the price-quantity pairs of the industry in which the
innovation occurs and of its subsidiaries and competitors will be affected
in ways that do not call for additional comment. Occurrence of innova-
tion must, as mentioned above, be located historically1 in every indi-
vidual case, a task necessarily preliminary to any thorough price-quantity
analysis. Recalling what has been said before and using traditional
concepts, we may now define the ultimate task of such analysis as follows :
given time series of actual price-quantity behavior, to resolve them into
five components, namely, movements along supply or demand curves,
shifts (including changes in form) of demand curves within invariant
indifference varieties (constant tastes), shifts (also including changes of
form) of supply curves within invariant production functions, autono-
mous changes of tastes, changes of production functions (innovation).
The task is difficult if numerical exactness is the goal. It is, though
laborious, not difficult if we content ourselves with a rough common-
sense approach. In particular, there is no difficulty in accounting for
the behavior of price-quantities in the prolonged periods of adjustment
that are sometimes induced by innovation and during which, many or
most firms continuing to produce at a loss for years, the survival of the
unfit creates that species of overproduction with which we are familiar.
Other deviations of price-quantity pairs of innovating industries from
the average of all commodities, which might easily be mistaken for spe-
cial cycles or trends, are not less easy to understand.
But if innovation consists in the introduction of a new commodity,
the entrepreneur finds himself, as we have seen, almost invariably in an
imperfectly competitive situation. In most cases his enterprise also
impinges on a sector in which imperfection prevails independently of it,
so that the little which, for our purpose, it is necessary to add to what has
been said on the subject in the second and fourth and in the historical
chapters, can be conveniently dealt with under this aspect. We then
meet again the phenomenon which we call stability of price if we approve,
and rigidity of price if we disapprove of it. Although there is little
reason to believe that this stability or rigidity has been on the increase
during the last fifty years, and that it really merits the attention which has
been paid to it of late, it is important for our purpose to understand what
it is that prevents so many prices of new as well as of old commodities
from participating in the cyclical fluctuations as prices under perfectly
competitive conditions would.
1 It follows, however, from our discussion in the third chapter that the occurrence of
innovation may be indicated by the behavior of marginal cost corrected for change in
prices of factors. The development of this tecbftique and corresponding fact finding are
urgent desiderata.
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 537
Traditionally sticky and authoritatively regulated prices do not,
from our standpoint, differ very much. That cycles, Kitchins in partic-
ular, will in many cases be missed by commodities and services that fall
in either class is too obvious to detain us. But three points deserve
notice — all of them relevant also to the examples that follow. First,
any industry or firm which asks a public service commission to sanction
an increase or a decrease in price, must know that the first is extremely
difficult to grant and that the latter is, for the same reason, likely to be
definitive. It requires spectacular emergencies — inflation may not be
sufficient — to "justify" an increase in the eyes of an invariably hostile
public opinion. This is an external factor. Second, it does not neces-
sarily follow that quantities will now fluctuate more than they would with
a more variable price. As it happens, many of the commodities in
question meet with a demand of low elasticity, which also does not much
shift in the course of cycles. It is often doubtful whether quantity
sold would, in an ordinary depression, be appreciably greater if prices
were promptly reduced. Third, in not unimportant cases demand,
without being inelastic, is such as to react only to price reductions so
considerable as to be normally out of the question without an important
innovation. Quite frequently, reduction will have sense only if the
intention is to tap new strata of consumers. But such a possibility is not
necessarily associated with any particular phase of the cycle, and may well
offer itself in prosperity.
We will next consider the case of a few-firm industry the units of
which have, through mere sluggishness of business spirit, for some time
been left in undisturbed possession of the field. The firms are assumed to
be no less sluggish themselves and to have settled down, but without
agreements, in an oligopolistic quasi-equilibrium buttressed by product
differentiation. Such an oasis, it is true, is not easy to find — most
instances that one might think of proving to be spurious on analysis —
but it will serve to illustrate a point we wish to make. Let prosperity
set in. The firms will enjoy brisker sales, but unless they are unable to
increase their output without incurring higher cost per unit, they will
not readily take advantage of the situation by raising their prices; for in
this case price cannot impersonally rise as does the price of wheat.
Somebody who can be identified has got to do the thing; and he risks
losing custom, even if the others eventually follow, which at least some
of them may not do at all, because they hope to conquer some ground
that they did not dare to invade initiatively. Similar considerations
apply on the downgrade. It should be observed that the very stable
prices that statistics will report while such a state of things lasts, are not
inflexible in the same sense in which a traditional or regulated or agreed
price is. They could vary at any moment and are chosen on strictly
538 BUSINESS CYCLES
rational considerations, with a view to what under the circumstances is
maximum advantage over time. We have in this case assumed sluggish-
ness, not in order to suggest irrationality, but only in order to exclude
the spirit of innovation and propensity to fight.
Irrational elements of course come in to help to petrify. In spite of
both rational and irrational elements, the situation is not likely to last
in a world pervaded by our process. But if existing firms are satisfied
with it, they may and often do attempt to peg it — which, for some of them,
implies readiness to forego such individual opportunities as they may
have reason to expect from the incessant revolutions wrought by that
process — and to adapt it by corporative action to changing conditions.
They often try to "rationalize" it by agreeing to outlaw "cross hauling"
— which, however, is not easy to recognize in the case of product differen-
tiation. They may get together in order to "educate" each other up to
an "ethical code."1 Or they may simply decide to prevent their resale
agencies from starting on a warpath of their own — one of the motives,
though not the most important one, for the movement toward Resale
Price Maintenance.2 They also may corporatively adopt a policy of
Basing Point or of Delivered Prices, as, for example, has been done in
this country by the steel, cement, lumber, paper, petroleum, beer, rubber,
glass, fertilizer, flour, and sugar industries. We are not here concerned
with the merits and demerits of these practices, into the discussion of
which enters so much "theory" that really deserves the quotes. But it
should be observed that, while some of them may work in the direction
of perfect competition and none of them necessarily spells rigidity of price,
yet all of them require machinery which is difficult and sometimes costly
to set into motion. We had examples in the historical chapters. Indus-
trial resonators are structurally affected by those policies, and failure of
prices to react to cyclical situations is, in a very obvious way, often
accounted for thereby.
If policies of that type issue in a cartel of the German kind, the work-
ing of the brittle monopoly that may be enjoyed by the industry in the
intervals between recurrent breakdowns reveals, first, the main rational
and purely economic reason for price stability in a monopoly situation.
1 Such attempts are sometimes unintentionally humorous. But their meaning is not
disposed of by the proof that some arguments used — such as the argument that total
average cost must always be covered — are erroneous. See, for an instance, E. C. Brown,
Price Competition in the Commercial Printing Industry of Chicago, Journal of Political
Economy, 1930. Notice, in particular, the Code of Ethics adopted by the United Typoth-
etae of America, p. 199. How many of us, and for how varied reasons, yearn for the
spirit of the Middle Ages!
2 Examples abound; see, for instance, E. T. Goether, Resale Price Maintenance in Great
Britain, Quarterly Journal of Economics for August 1934. This paper has the merit of
bringing out very clearly that such attempts are like canoes trying to live in a stormy sea.
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 539
It has been mentioned above in another connection: a monopolist, if he
faces a demand which, within the useful interval, either is always insensitive
to price or becomes so when shrinking in depression, will have no motive
to reduce price, while in perfect competition it would always fall. This
is so with many highly finished consumers' articles of small importance
in the household budgets and with some important producers' goods that
individually contribute but little to the total cost of any product. But
it should again be emphasized that the departure of such prices from
the cyclical pattern which would prevail under conditions of perfect com-
petition does not, in the case envisaged, imply corresponding deviation in
quantities, which do shrink, no doubt, but not primarily because prices
stay up and not much more than they would in the absence of monopoly.
As far as this goes, therefore, certain well-known propositions about
price rigidities that intensify depression do not follow, and there is some
justification for that attitude taken both by business interests and
politicians which in our days crystallized into the NRA legislation.1
Second, that tendency of monopoloid situations to produce stable
prices is reinforced, even if there is no cartel or monopoly in a strict sense,
by the practice of providing capacity for cyclical peak demands . As stated
before (fourth chapter), this practice may result in the prices of the most
cyclical industries becoming the least cyclically variable of all. This
phenomenon would also be impossible in perfect competition. It does
intensify fluctuations in quantities but rather mitigates than intensifies
cyclical difficulties. Third, the rational but extraeconomic considerations
that have been mentioned above greatly influence a capitalistic combine
which knows that it is unpopular, that any reduction of prices will be
looked upon as a proof of past exploitation, and that it may be impossible
or difficult to increase price again. The policy of letting sleeping dogs lie,
therefore, frequently recommends itself. This applies to any monopoloid
situation, but in the case of a cartel there are, fourth, the difficulties of
taking action in the face of divergent interests of members. In Germany,
for instance, it was often the case that the biggest and most efficient
firms opposed increases and sponsored decreases of prices and that the
smaller and less up-to-date firms clamored for the former and fought the
latter, bitterly accusing the big men of being "apostles of abstemious-
ness" (the writer is trying to convey the meaning of Maessiglceitsapostel).
There were many reasons for this. One of them, however, was that the
big firms did not relish the idea of having to back a policy which pre-
1 The above must not, however, be interpreted as a defense of any particular policy.
All it means is that there are cases in which measures of a monopolistic flavor cannot be
condemned on traditional antimonopolistic grounds. See, however, Professor Karl
Pribram, Controlled Competition and the Organization of American Industry, Quarterly
Journal of Economics for May 1985.
540 BUSINESS CYCLES
vented them from using their powers to the full and tended to preserve
economic lives the necessity of which they were unable to see. Under-
standably, this often spelled deadlock,1 with the result that, no decision
being arrived at, prices did not change at all or changed with a perfectly
erratic lag.
On the whole, readiness to combine is more marked in Kondratieff
downgrades than in Kondratieff prosperities. (For examples see Chap.
VII, especially sec. D). But there are many exceptions to this general-
ization. It seems that the structure and conditions peculiar to the
individual industry count for more than does the general business situa-
tion. Nor is this astonishing. When things look bright, one of the
strongest motives may be lacking for firms to take shelter, but others are
gaining in importance and there is less motive for breaking away in
despair — vice versa in depression. Now since an organization that
really constitutes anything at all approaching monopoly, will, as we have
seen, powerfully influence the price-quantity pairs of the industry, its
foundation or breakdown will dominate their statistical picture and may
even blot out the influence of cyclical fluctuations altogether. Examples
abound and need not here detain us. There is nothing in this to modify
our theory of cycles or the expectations that follow from it.2
If a cartel breaks down, or if a quasi-equilibrium among oligopolists
is disturbed, it does not follow that we may now substitute the competi-
tive schema to the one that applied before. On the contrary, there will
be what we call a Disorganized Market. We shall observe those moves
and countermoves, which have little relation to costs and may have
almost as little to the cyclical situation, particularly if, during the rule of
a cartel, excess capacity has developed: bellum omnium contra omnes.
1 So did other things which are very real, though not always easy to establish beyond
doubt. There is a committee-room etiquette, departure from which, even if unintentional,
may be felt to be an outrage, driving the offended person into restive ways which we should
unhesitatingly call feminine if we observed them at a tea party. There are armchairs and
ordinary chairs. A history of the secrets of big business is much more likely to bring out
the importance of this class of fact than to reveal dark deeds and profound plans.
2 Responsibility for a large fixed investment may — though mostly it does not — account
for a wish to reduce risks by combination; necessity of committing oneself to large fixed
investment may account for the existence of a monopoloid situation. In either case, there
is a connection, at one remove, between size of overhead and some motives for a policy of
price stabilization. But it is perhaps not superfluous to say that there is no other. The
mere fact of relative or absolute largeness of overhead would not, in itself, make prices more
stable than they otherwise would be; it would, if anything, tend to make them less so. The
contrary opinion would be nothing but an error, if it did not acquire some importance from
the fact that business practice often shares it. But observe that the presence of the error
in the practitioner is not proved by the facts, either that he uses the fixed cost argument in
justification of his refusal to reduce prices in depression to marginal costs or that actually
he does not do so.
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 541
The American paper industry and, to a certain extent, the American
shoe industry in the twenties of this century, or, at various junctures,
the German cement industry, may be cited as instances. But such a
situation will not always be faithfully reflected in the behavior of price-
quantity pairs; for there will be rebates, credits, and concessions in other
conditions of supply. Changes in quality will replace changes in prices
and make it difficult to follow up changes in quantities. In fact, a
situation like this can persist for some time without affecting the quota-
tions from which statistics are derived. This is a case of spurious or
statistical stability.
This analysis also applies to the case in which it is a new commodity
that supplies the upsetting impulse, although the economic meaning of
the ensuing competitive struggle is profoundly different, the position of
competitors being not unstable merely, but inherently untenable. Only
two points call for additional comment. First, for a considerable time
during which the new article is vigorously gaining ground, its price as
well as its quantity may be very little sensitive to cyclical fluctuations.
Demand may go on shifting upward through several consecutive depres-
sions of the Kitchin, possibly even of the Juglar, and there may be no
reason for the innovating firms to change their prices. There are many
instances of this in such fields as the motorcar (carrying with it gasoline,
some rubber goods, nonshatterable glass, and so on), electrical apparatus,
harvesting machinery, rayon, alloy steels, motion pictures, and other
industries. Needless to say, such behavior is for us anything but con-
trary to expectation, however much it may deviate from average behavior.
Here, again, stability of price is in a sense spurious. For prices are stable
not primarily because they are kept so, but because opposing forces,
acting upon price and quantity make them so: another instance to show
that prices which do not change for a time are not, ipso facto, rigid in
our sense.1
The second point refers to those situations in the cyclical process of
evolution in which entrepreneurs temporarily enjoy what, on a previous
occasion, has been called an acceptable approximation to straight
monopoly (Chap. II, sec. F, III). In such situations, several courses are
open to them. For instance, they may decide to make hay while the
sun shines, to use this hay to write off their plants to $1 as quickly as
possible, and accept defeat without struggle when their demand curve
crumbles. It should be observed that in such truly monopolistic cases
it would, in general, precisely not be to the interest of a seller to keep his
prices stable through changing situations. But as a rule an entrepreneur
has no such chance and no such intention. He must build up demand
1 See, however, V. A. Mund, Prices under Competition and Monopoly, Quarterly Journal
of Economics for February 1984.
542 BUSINESS CYCLES
and then defend the ground conquered against the attacks of competitors
for whom he fatally paves the way. Hence, he is neither before nor after
success in a position to behave according to the schema of the classical
theory of monopoly. In those cases in particular, which have steadily
grown in importance ever since Huntsman produced steel in the dark of
the night, by the work of his own hands, in which the entrepreneurial
achievement consists in or presupposes the creation of facilities for mass
production, demand can be built up and the ground conquered be
defended only by selling cheaply from the start and never raising prices
afterward. Visualizing the price that in the average of good and bad
years will attain both ends, and being able to produce at costs which
that price will more than cover is, in those cases, the main requisite of
success. Theoretically, such a price need not be inflexible downward,
but it is easy to see that under the circumstances any reduction, unless
based on quite permanent conditions, must be a dangerous step to take.
Hence we find, mainly in the field of highly finished consumers' goods but
also in the field of tools and machinery, so frequently prices that according
to all ordinary standards are extremely rigid — the price becoming part of
the individuality of the commodity and the firm being practically com-
mitted to it. But the point to be made is that this kind of rigidity
differs from, and does not entail the same cyclical consequences as,
rigidity in the usual sense. The latter is defined by, and its effects turn
on, deviation from the price that would at any moment be ideally — never
mind now whether in a monopoly or competitive sense — adapted to the
general situation.
A price of the kind envisaged will not conform to that standard, but
it will, since it has been fixed with a view to an expected sequence of
situations, come in case of success — which of course is a matter of a lucky
throw — much nearer to it than a price fixed, say, by compromise such as is
characteristic of a cartel or by some rule dictated by a public service com-
mission. It is also not true that in these cases, as a well-known slogan
has it, "quantity adapts itself to price instead of price to quantity,"
because considerations about future quantities of output precisely hold
first place in determining the decision. Moreover, it must not be for-
gotten that any new article, however inflexible its price may be, is still
a tool of flexibility in the system as a whole, by virtue of its impact on
the preexisting price-quantity structures in its field. Finally, if adjust-
ment becomes necessary after all, there will be motive to effect it through
change in quality or, if that be impossible or undesirable, by offering
another type or many other types of product at different prices. Statis-
tics will then, if the old type is not discarded, record absolutely rigid
prices for both the old and the new type, while to all intents and purposes
there is — an inhibited, no doubt, and discontinuous — flexibility and the
PRICES AND QUANTITIES OF INDIVIDUAL COMMODITIES 543
usual inferences from price rigidity are largely unwarranted, though
of course not wholly so. This is particularly clear if several similar
articles are offered from the start — "a car for every purse," for instance —
when no actual change of price is necessary at all in order to produce
practically all the effects of prompt variability.
Thus, analysis of the nature and sources of the various kinds of price
rigidity we observe and of that monopolistic or oligopolistic strategy
which, intentionally or unintentionally, rationally or irrationally, is
responsible for some of them, hardly lends support to the ideas many
students entertain about their importance or, as some would say, growing
importance for the cyclical mechanism, particularly, their dislocating
effects on the rest of the system in depression. There is less genuine
rigidity, and what there is of it is less dislocating, than is widely assumed.
We should not, however, run into the opposite error. The very fact that
every concern which has any opportunity to do so strives to secure a posi-
tion that looks as monopolistic as possible — by advertising, differentiating
its product, trying to control competitors (electric power and light com-
panies, for example, tried to get control of gas companies; American rail-
roads tried to buy up trolley lines), warding off competition by preventive
attack — is in itself sufficient to prove that such positions are not valueless.
But we have seen in our historical sketch that their value consists,
rather than in any power to follow a long-run policy of restriction of
output, in the facilities they afford for steering safely through difficult
situations and for undisturbed planning. And the difference the existence
of such positions makes is not so much a difference in ultimate results
as a difference in the way by which they are reached. The latter is, to be
sure, quite sufficient to upset our expectations as to the behavior of price-
quantity pairs in the short run, i.e., practically in the Kitchin. But it is
not sufficient to upset the working of our process.
CHAPTER XI
Expenditure, Wages, Customers'
Balances
A. Some Propositions about Money. — Propositions about money
have been introduced already, and others will be when necessary. Most
of them are to be found in Chaps. Ill, VIII, and XIII. Still others
will be assembled in this section. Since it is not possible to deal in
this book with the general theory of money to which these propositions
pertain, they must here appear in an incomplete and unsatisfactory form
and detached from the background on which alone they could acquire
their full meaning.1
1. Money may be, and in practice mostly is — at least historically —
linked to some commodity. But it never is a commodity and never
satisfies wants in the sense in which commodities do. If we nevertheless
attribute utility to it, this utility is derived from that of the commodities
we actually buy, or could buy, with it and hence presupposes given prices,
or ratios of exchange between money and commodities. Any attempt,
therefore, to deduce these prices from marginal utilities of money and
marginal utilities of commodities, in the way in which we may deduce
exchange ratios between commodities from their marginal utilities,
involves circular reasoning. This is so even if the monetary unit consists,
for instance, of a given quantity of metal that can be freely coined and
melted without cost or loss of interest. For although in this case the
exchange value of that quantity of metal in its monetary use can never
depart from its exchange value in its industrial uses, the former cannot
fundamentally be explained by the latter : the closing of the mints to the
public would suffice to bring out the fact that money has an exchange
value of its own, which, therefore, as long as coinage is unrestricted,
"determines" the exchange value of the money commodity in its indus-
trial uses quite as much as it is "determined" by it. Unfortunately we
cannot enter into the discussion of the various devices by which theorists
have labored to avoid the implications of this. It is hoped, however, that
our argument is clear as it stands. If so, we may draw the conclusion
that any kind of linking of the monetary unit to the unit of a commodity,
1 The writer hopes to provide that background and to develop the theoretical structure
of which these propositions are fragments, in his treatise on money.
544
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 545
whatever its practical merits in guaranteeing the value of money may be,
is logically nonessential and subjects the functioning of the monetary sys-
tem to an additional condition which is extraneous to the meaning of
money.
2. The fact that money is not a commodity explains what otherwise
would be inexplicable, namely, that claims or titles to money (however
defined) may serve the same purposes as money itself. This is the funda-
mental explanation of the possibility of "credit creation," as well as the
reason why it is so easy to create "near money"1 and so difficult to
prevent the creation of it. No such thing can occur in the case of a
commodity.
3. The same fact underlies the phenomenon of "velocity of money,"
which, similarly, has no analogy in the world of goods and services.
This phenomenon is not correctly described by calling money a durable
good which can serve many times. Money serves merely as a counter,
which, within technical limits, can turn up any number of times during
the game. These limits are essential, however. The "periods," i.e.,
the spans of time which it takes the monetary units to complete their
circuit, are the fundamental facts about monetary circulation.
We have to distinguish three concepts of velocity: (a) the sum total
of all transactions in terms of money divided by checking balances plus
money outside of banks; (6) consumers' plus producers' expenditure by
balances plus money in circulation; (c) consumers' expenditure by
balances plus money in circulation. The last, the so-called income
velocity — we may call it net income velocity to distinguish it from gross
income velocity equal to total income by balances plus money in circula-
tion— is the most relevant figure of the three; the first is, in itself, mean-
ingless, though it must sometimes serve as the only indicator we have
of the other two.
More important, however, is another distinction, which may be
applied to any of the three velocity figures just mentioned. In a station-
ary state, velocity would be largely determined by the institutional
arrangement of payments within the period of account. Suppose the
economic process to be organized in such a way that all firms buy produc-
tive services from households, say, on each Saturday, and all households
buy consumers' goods from the same firms and for the same amount of
money, on each subsequent Monday. This being all that happens,
income velocity per year is 52. We may go a step further, however,
and assume that incomes are paid out and spent concurrently all through
1 The writer supposes that that term has been coined by way of analogy with "near
beer." If so» it does not express the whole truth. Near beer did not serve so well as beer
and it had to be produced. It is, also, not quite correct to say that "credit" serves as a
"substitute for money."
546 BUSINESS CYCLES
the week in small installments, and that the sequence of everyone's indi-
vidual payments and receipts is random. This case will, of course, give
a different velocity and, moreover, display an element absent in the first
case. Households as well as firms will now hold an element of cash
specifically intended to provide for the occurrence of unfavorable
sequences. Subject to this new proviso, however, there is still a deter-
mined velocity measuring the way through economic space of every unit
of money spent — and all units that are spent are assumed to be promptly
spent, for there is, in this setup, no object in withholding them. We
shall use the word efficiency for this kind of velocity, or confine the term
velocity to it. For it the classical assumption of constancy or slow and
independent variation is approximately true, not only in the stationary
state, but in all cases, that of extreme inflation alone excepted.
As soon, however, as we leave the precincts of the stationary process,
in order to deal with changing business situations, we meet besides effi-
ciency a phenomenon which is entirely different from it, although it
influences velocity figures similarly. We find that people sometimes do
withhold money they intended, and still intend, to spend; and that,
whereas before spending was a matter of course, the question whether or
not to spend at any moment becomes a question of policy for everyone,
obviously important for the picture the monetary process will present.
While efficiency refers to the velocity of any unit that is actually sent over
its path, we now find another component of the velocity figure which
refers to the proportion of existing units so sent. We shall call it Rate
of Spending, or simply Spending (see Chap. Ill, sec. A), and shall put
its equilibrium (stationary) value as equal to unity. Obviously, this is
the cyclical variable within total velocity.
4. Since certain claims to "money" serve, within wide limits, the
same purposes as legal tender itself, it is not only necessary to include the
existing amount of such claims in the total quantity of money (typical
cases: bank notes, deposits subject to check), but also evident that
the very concept of quantity of money becomes doubtful. It is, in fact,
impossible to speak of the quantity of " money " in the sense in which we speak
of the quantity of a commodity.1 We therefore stand to lose what, as we
have seen in the chapters on price level and prices, is nevertheless a
necessary element in the determination of values. What is to fill this
breach in the theory of money is a problem which cannot be attacked here.
In any case, however, we cannot here consider quantity of " existing "
or "circulating" or "available" money as an independent variable,'
because, although it varies in function of some elements that may, in the
1 Another consequence is that the distinction between velocity and quantity becomes
blurred. K. Wicksell, for instance, treated the issue of bank notes as a means of increasing
not the quantity but the velocity of money (the banks' reserves).
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 547
sense of the old quantity theory, be looked upon as data, it also varies in
response to other variables of our process, entrepreneurial activity in
particular. It has been pointed out before that many modern authors —
most of them adherents of the investment theory of banking (Chap. Ill)
— try to meet this situation by what may be interpreted as a revival of
the quantity theory. They figure out what the technical maximum of
credit creation is within a given institutional frame — legal or traditional
rules about reserve proportions and the like — attribute to banks a
tendency to maintain that maximum of customers* balances and thus
construct a quantity of money which is held to act as such on the economic
process (and to have a causal significance) in much the same way as the
quantity of money in the quantity theory sense was by older authors held
to act. In Chap. XIII we shall see how far from satisfactory any such
mechanistic theory of banking is. Here it is sufficient to state that, even
disregarding the fact that that institutional frame is not independent of
the evolutionary process, the maximum alluded to does not constitute
the supply but at best the limit of the supply of balances, and that the
amount banks are, in a given situation, willing to supply cannot be
explained by considerations at all analogous to those that are applicable
in the case of the supply of a commodity.
5. Money not being a commodity, the traditional apparatus of supply
and demand cannot be applied to the solution of the problem of money
prices of commodities and of the price levels (see Chap. VIII). An
exchange of money for commodities is not the same phenomenon as an
exchange of one commodity against another. This is still clearer on the
demand side than it is on the supply side. We may, with some qualifica-
tions, speak of a demand for money in the money market. But there
is no sense in speaking of a demand for money displayed by sellers in a
commodity market. Demand for money carries, however, still another
meaning: it may mean the wish to hold stocks of money or balances.
This Walrasian idea of an encaisse dSsirSe, which reappears in Marshallian
analysis and has recently, it seems, entered upon a new lease of life, is one
of the least valuable elements in the great Frenchman's mighty structure.
It is harmless only in the analysis of stationary states, although even there
it implies a misrepresentation of facts. If people get their "incomes"
each Saturday and spend them on consumers' goods each succeeding
Monday — transactions between firms being excluded — then the money
will lie about in the vaults of firms from Monday to Saturday, not
because there is any demand for cash holdings, but because the institu-
tional arrangement so wills it. But the idea becomes misleading if we
leave the stationary case. If we see someone displaying a wish for bread,
this is a clear-cut fact carrying its explanation in itself and fit to be used
in order to deduce the explanation of other facts. But if someone dis-
548 BUSINESS CYCLES
plays a wish to hold cash, this in itself means nothing at all. All the
value of the observation lies in the circumstances that induce that wish —
all the theory of the fact and its consequences turns upon those circum-
stances— even if there is such a wish. But generally there is none. A
man may, for example, hold a supernormal amount of cash, not because
this is any good to him, but simply because his and other peoples' actions
happen to produce that result, which in itself is not one of the objects he
wishes to attain by those actions; it may even be a disagreeable by-prod-
uct of them. All explanations which start with the famous adage: "If
people choose to hold • • • " are ipso facto condemned.
6. A general remark may conveniently be inserted here. One of the
first tasks that confronted scientific economics in the early stages of its
career was to fight certain popular views about money which, however
understandable or even defensible they may now seem to the historian,
were in fact largely erroneous. During their campaign against "bullion-
ist" and "mercantilist" exaggerations of the importance of the role of
money, economists were naturally driven to construct a body of doctrine
in real terms alone. They tried to draw away the "veil" of money in
order to describe the process of the production and consumption of
wealth. This effort, although highly meritorious at the time, was bound
to fail in the end. Economic action cannot, at least in capitalist society,
be explained without taking account of money, and practically all
economic propositions are relative to the modus operandi of a given
monetary system. In this sense any theory of, say, wages or unemploy-
ment or foreign trade or monopoly must be a "monetary" theory, even
if the phenomenon under study can be defined in nonmonetary terms.
This is increasingly being recognized, and the fact that it is must be listed
among the major improvements our analytic apparatus has undergone
during the last 20 years or so. But we seem unable to draw away from
an old error without running into the opposite one, which in this case is
older still. That economic analysis cannot — in the sense in which for
instance Boehm-Bawerk thought it could — abstract from money is a
truth which is useful only if supplemented by the other truth that
monetary processes never carry their explanation in themselves and can-
not be analyzed in monetary terms alone. And recognition of the fact
that, in its fight against mercantilist views about the causal role of money
in economic life, classical doctrine went much too far must be supple-
mented by the other fact that in fighting them it also performed a service
of which we again stand in need today.
B. System Expenditure (Outside Clearings). — The changes in
monetary expressions and monetary quantities that occur in the course
of the cyclical process of evolution either reveal themselves in, or are
brought about by, changes in the totals of business and household
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 549
expenditure, which are — and so far we agree with the most monetary of
theories of the cycle — the most important immediate conductors, as well
as the most obvious effects, of changes in the general complexion of busi-
ness situations. In a general theory of money it would evidently be
appropriate to consider producers' and consumers' expenditure as
mutually interdependent: producers' expenditure expands and contracts
in function of consumers' expenditure exactly as consumers' expenditure
expands and contracts in function of producers' expenditure, that is to
say, in response to the changes in income paid by firms to households.
But for the particular process with which we are concerned here, it is
more helpful to recognize that the fundamental impetus comes from
producers' (entrepreneurs') expenditure and that households merely
react to it.
It is necessary to use this proposition with care. Instances of the
reverse relation are not lacking, even within our process: that reaction
of the business world which makes up what we have called the Secondary
Wave is partly reaction by firms to increased consumers' expenditure.
But inasmuch as these phenomena can be traced back to the impulse
given by entrepreneurs' expenditure made in the course of carrying out
new combinations of factors, they may also be considered as consequences
of that prime mover. Not all incomes, moreover, are paid out by firms,
nor do all that are, vary with the sum total of producers' expenditure.
Exceptions must be made for them. But, with these qualifications and
in this sense, it is still broadly true that household expenditure varies in
function of producers' expenditure, although, of course, not simply in
direct proportion to it. Common sense and common experience will
probably find little fault with the simplifying generalization that pro-
ducers' expenditure is, within our process, the active element of total
expenditure in the system (consumers' plus producers' expenditure) or,
to use a term due to Mr. C. E. Thomas, of system expenditure. We shall
come nearer to facts if we assume this quantity to vary, not only in func-
tion of producers' expenditure, but also in function of its rate of change,
which takes care of the most common type of anticipations as to the imme-
diate future: producers' expenditure molds consumers' expenditure, not
only by supplying households with money to spend, but also by shaping
their willingness to spend; and this willingness largely depends on the
rate of change of income streams which prevails at the moment.1 The
1 Even the relation, per period of account, between firms' expenditure and households,
receipts is, owing to lags and to payments between firms, not simple or invariant, let alone
one of identity. Still more important is it to observe that, although there is no doubt about
the direction of the influence of household's receipts on household's expenditure on con-
sumers' goods, there is plenty of doubt about the form of this function also. It may vary
widely as between nations and situations. The French peasant and small bourgeois, as well
as — perhaps — the "Scotchy" New Englander of old, may provide instances of a limiting
550 BUSINESS CYCLES
question of introducing a lag — which would in any case be a short one —
need not be touched at our stage of approximation.
The most comprehensive series of expenditure is the series of total
debits to banking accounts. In the United States this total may be said
to approach the sum of all monetary transactions, although an amount of
transactions still remains unrecorded which it is hardly possible to esti-
mate accurately and which of course varied greatly in the slow evolution of
banking habits during 150 years before the war. It should be added that,
exactly as some monetary transactions fail to show up in debits, so many
case in which increase in receipts entirely fails to induce an increase in consumers' expendi-
ture. Again, sudden increase in monetary receipts, due, for instance, to the incipient stages
of inflation, may sometimes also fail to produce a correspondingly great effect — that is why
prices rise less than balances or quantity of fiat money during those stages or why balances
themselves rise less than in proportion to government spending, the new access of wealth
being partly applied to the repayment of debt. But opposite cases seem to the writer,
under ordinary circumstances, to be much more frequent and important. The working
hypothesis which we shall mainly rely on is, hence, that consumers' expenditure increases
more than proportionately whenever household receipts increase and decreases more than
proportionately — particularly if durable consumers' goods play a great relative role in
consumers' budgets — whenever household's receipts decrease. This comes practically
to saying that there is a general tendency for the average household to outrun its "income"
in all phases of all cycles excepting depression, in which, however, forced dissaving and
subsidies counteract the effect to some — historically variable — extent. This hypothesis
rests on the following observations, which however, it will be observed, are all drawn from
"modern times" — roughly, from the time since 1900.
1. The writer thinks that he has indeed observed a type of industrial family (and, in
some countries, of peasant family, as stated above), though much more clearly in Austria,
England, France, and Germany than in this country, which keeps to a very rigid style of
life, mainly expressed in its family houses and the way they are run, and varies its expendi-
ture but little as between prosperity and depression. The vast majority of all classes
behaves differently. One stratum (witness the strongly cyclical character of the jewelry,
theater, and hotel trades) obviously spends the speculative and other temporary gains —
the former are not income in our sense — of prosperity on consumers' goods. The wage-
earning stratum seems to the writer to display the same bent and to try in times of rising
wages to spend "all it can." This is true even for noncyclical increases of income, such
as periodically occur in the salaries of state employees; in Germany, at least, they seemed to
spend more than the increase each time their salaries were augmented.
2. All those impressions are supported by the growth of the practice of buying on
installments, although this practice would not, in itself, suffice to establish them.
8. It is a fact that the general net indebtedness of households increases in times of
rising incomes, though this fact can be fully proved for postwar times only.
4. The presence of a tendency in the vast majority of people to spend on consumers'
goods whatever they can, excepting depression, is suggested — though again not strictly
proved — by the monthly and weekly rhythm in retail trade, (cf. Gruenbaum, Umsatz-
schwankungen des Einzelhandels, Sonderheft 10 of the Berlin Institute's Vierteljahrshefte,
1928.)
5. The contrary impression, which so many economists seem to have, can be traced
to the one item to which the above clearly does not apply, viz., profits in our sense.
We shall return to the subject presently, and again in Chaps. XII and XIV.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 551
commodity transactions fail to give rise to monetary transactions —
especially the "transactions" that consist in producers', particularly
farmers', "selling" part of their own produce to themselves. This debit
series is, however, not available for the prewar time. Instead, we have
the series of bank clearings, which must be expected to differ from the
true figure of debits that it is supposed to indicate, and even this we have
only for this country and England. But the American figure has been
shown to display so close a relation to the figure of debits since the war,
that we may with some confidence use it. In countries and for times in
which banking is highly concentrated the situation would, of course, be
less satisfactory.
Apart from the changing relative importance of the amount recorded
and various difficulties of a technical nature into which we cannot enter
here, it has to be recognized that the total is, as such, a composite of very
little significance. Debits arising out of transactions between banks have
a special character and should be excluded, as should — for some purposes,
though not for others — debits to public accounts. Moreover, the total
records every charity and every tax payment, as well as disbursements
from incomes which are not paid out by firms. However, some of these
elements are not important enough to distort the picture; others are as a
matter of fact excluded. What seriously impairs the value of the clear-
ings series is the impossibility of separating stock exchange and real
estate transactions from the rest. All we can do about this is to adopt
the familiar distinction between New York and Outside Clearings,
although the former contain all the transactions of the world's greatest
industrial and commercial center and the latter, a very considerable
element of stock exchange business and of speculative transactions in
general. Similarly we must, in the English case, pin our faith to the
distinction between town clearings at the London clearinghouse and
provincial clearings plus country clearings at the London clearinghouse.
Bearing in mind, then, that they really measure something different
and, in particular, that they record the value of every element of every
commodity as often as it changes hands, against payment by check,
in the course of production and trade, we must resign ourselves to using
Outside Clearings1 (Mr. Frickey's series mainly) to indicate variations
1 Series are available monthly for a varying number of cities, which increased to 159,
from the Financial Review, the Public, the Commercial and Financial Chronicle, and other
sources. Mr. Snyder's compilation goes back to 1875 (Journal of the American Statistical
Association, 1924). He corrects by his general price level. Mr. Frickey's monthly series
for seven selected cities (Baltimore, Chicago, Cincinnati, Cleveland, Philadelphia, Pitts-
burgh, San Francisco) is the only one to take account of the fundamental difficulties of
this material, consisting in the effect of new clearinghouses emerging from time to time,
before 1903 (see E. Frickey's paper in Review of Economic Statistics for October 1935).
Some experimentation has quickly shown it to be, because of this and other virtues, the
552
BUSINESS CYCLES
"NEfDEPOSIT:
INVESTMENTS
MINUS
OUTSIDE N.Y.C,
f,''\'C( EARINGS
in system expenditure. Some idea of the extent of the risk we run when
reasoning as if outside clearings represented the dollar volume of physical
production may, however, be derived from inspecting Chart XXIII.
Whatever confidence we may place in the evidence it presents, it certainly
goes some way toward allaying extreme apprehensions. The covariation
between Outside Clearings and Physical Output of Industry times the
Price Level is, given the statistical in-
dependence of the respective materials,
in itself an interesting result. Possi-
bilities of very simple "theories" seem
to open up at this point.
System expenditure makes a natural
and systematic series of the "velocity"
(or rate-per-time) type, which is ob-
viously cyclical. It is a primary but
consequential phenomenon. There is
no need to stress that it is primary : in-
crease in system expenditure is not only
one of the most obvious elements in
the picture of any prosperity, but also
the most important factor in produc-
ing the symptoms that we associate
with prosperities. All the more neces-
sary is it to emphasize that it is neverthe-
less consequential: no antecedent or
initiative increase in system expendi-
ture— induced for instance by some
chance event or by the initiative of
banks or governments — is required in order to start the extrepreneurial
activity that propels the system away from any preexisting neighborhood
of equilibrium. This activity directly implies and indirectly induces all
that additional spending we observe in prosperities, but the innovations
themselves are independent of it in the sense that they are profitable at
the system expenditure in the preexisting neighborhood.1 If they are
J_
/MANUFACTURING
/AND MINING X PRICES
_L
1890 95 1900 05 1910 13
CHART XXIII.— United States (see
Appendix, p. 1059).
best one to use, and, with one exception, it has been primarily used by the writer in his
work on system expenditure.
1 See the analogous statement about the price level (Chaps. IV and VIII). There is,
thus, so far neither necessity nor room for the assumption that either "spending" itself
or the sphere of money and credit which provides the means for spending harbors any
cyclical mechanism of its own. If our analytic schema be accepted, all the features of the
behavior of spending and of credit in a normal cycle can be accounted for without assigning
to them any but an adaptive role. Since the above contains our solution of a much-debated
problem — it might be termed "the riddle of spending" — and since the opposite view is, in
one form or another, almost universally accepted and not without influence on policy, the,
reader is invited to recall once more the premises of the statements in the text.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 553
not, they are "maladjustments," which may have to be liquidated sub-
sequently. Similarly, system expenditure slackens or may even abso-
lutely decrease in recession, and this not only constitutes one of the
most regular features of recessions but also produces others. But it is
itself induced by the slackening of innovating activity and by autodefla-
tion, both of which again are independent of it in the sense that they
come about without spending or credit taking the initiative in calling a
halt. No doubt there may be, in any particular situation, also a pull
of the monetary bridle. But (see Chap. IX) such a pull is not necessary
in order to account for the occurrence of an "upper turning point," which
is perfectly understandable without it and does not, in system expenditure
any more than in anything else, constitute a distinct problem calling for
special solution and requiring the introduction of facts extraneous to the
mechanism of innovation. In the midst of increasing rates of expendi-
ture, and in the absence of all extraneous limits — monetary or other —
business could and would recurrently start deflating itself by virtue of the
working of that mechanism, and this indeed means decrease in rates of
increase or absolute fall of expenditure, but does not presuppose that it
occur from other reasons. Failure to realize this leads to many blind
alleys.
Expectation for the cyclical behavior of system expenditure seems to
follow arithmetically from the behavior of price level and output.
Expenditure is, however, not only a monetary expression, but also a
monetary quantity that exists and behaves as such, so that our expecta-
tion about it must be formed independently. In the "pure" model it
would rise in prosperity and fall to its previous amount during liquidation.
In the four-phase cycle, expectation for three out of the four phases is too
obvious to detain us, but for recession it is necessary to recall that owing
to the facts, first, that entrepreneurs* repayments do not actually go to
the length of eliminating their debts and, second, that other borrowing
partly, wholly, or more than wholly1 replaces entrepreneurs' borrowing,
our expectation loses its definiteness. All we can say is that system
expenditure will increase in prosperity more and in recession less than
total output, although we may also hazard the guess that it will increase
in the recession of every cycle at a smaller rate than — due attention being
paid to the simultaneous phases of the other cycles — in the preceding
prosperity. The resulting trend is, to be sure, still due to the working of
our process, but it nevertheless differs in nature from the result trends
which we observe in price level or output. These express fundamental
properties of the capitalist system and would be present even in a world
1 That will depend on the speed of the process of repayment and of the process of exploi-
tation of the new investment opportunities opened up by the innovations of the preceding
prosperity, and on how this "pushing into new economic space" is being financed. This
is not without some suggestions as to regulative and remedial policy.
554
BUSINESS CYCLES
conforming to the pure model, while the result trend in system expendi-
ture would be absent in such a world and enters our picture only at the
level of the second approximation.
Such as it is, that result trend is of course but imperfectly rendered by
the descriptive trend observable in our graph. The development of
deposit banking, the change in habits of payment incident thereto, and
all the changes in the institutional framework of the currency and in the
0.90
080
%
5 0.70
>°
10
-0.50
c
I 040
0.30
020
£+10
I -10
/ TREND FITTED BY LEAST SQUARES:
2>0.6l747 + O.OI780658!3t
(ORIGIN I896t)
^LOGARITHMS OF
HILADELPHIA CLEARINGS
(ANNUAL TOTALS)
I
DEVIATIONS FROM
IDEAL FORM OF COMPUTED PERIODS:
(12.9 YEARS AND 4 2 YEARS)
\m 1880 1882 1884 1886 1888 1890 1892 1894 1896 1898 1900 1902 1904 1906 1908 1910 1912 1914
CHART XXIV.— (See Appendix, p. 1060.)
rate of gold production also help to shape the latter. It must be recog-
nized, moreover, that the effects of all these factors are clearly not addi-
tive and that the effects of changes in the quantity of legal tender, in
particular, cannot be dealt with in the way which would have seemed —
and still seems — the obvious one to adherents of the quantity theory in
any of its primitive versions. The circumstance that none of those fac-
tors is independent of our process may perhaps be pointed to in explana-
tion of the fact that, barring times of war inflation, the behavior of our
series is, nevertheless, quite satisfactorily accounted for by our model and
that it does not obviously indicate any other than a passive role of money
and credit.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 555
Chart XXIV is presented for the sake of displaying an ingenious
method rather than for the results that can be gathered from it. In order
to apply Dr. Georgescu's method (Chap. V), it was necessary (the other
two differentiations which would have had to be added would have
accumulated errors to a dangerous extent) to eliminate a descriptive trend
fitted by least squares. The reader is invited to record and weigh this fact
against what has been said1 about this practice in the course of our con-
siderations on statistical method (Chap. V). Moreover, the material
covers too short a span to allow any formal method to show the Kondra-
tieff, or even to give reliable results as to the precise periods, or any precise
properties, of the shorter cycles. However, the hypothesis of a single
cycle yielded a period of 10M years, which, the writer holds, is strong
evidence both of the presence of a cycle of about that length and also of
its general properties' conforming to our expectations. The second step
of this analysis turned out less satisfactorily. When the material was
analyzed on the assumption of the presence of two cycles, the resulting
periods were 12.9 and 4.2 years respectively. We should not take it too
tragically that these figures exceed substantially what we consider to be
the average periods of the Juglars and the Kitchins. The fewness of the
cyclical units of both kinds may account for that, and the value of the
evidence for the presence of two cycles of about these periods is not
greatly impaired thereby. But it is more serious that the introduction
of the hypothesis of two cycles does not improve the picture of the
phenomenon yielded by the single-cycle hypothesis. The "fit" is some-
what improved, but not much.2
Turning to Chart XXV, we will first of all note the fact adverted
to above that not only the properties of the clearings series to be
expected from our model do actually show, but also that it behaves much
as if no other influences than those embodied in our model had acted
upon it. For 1875 to 1913 this is actually more nearly true than for any
other period within our range of vision. But we should not have been
surprised if such outside disturbances as nevertheless occurred had
asserted themselves more clearly. Detailed analysis no doubt reveals
some of them. The steepness of the ascent from 1878 has, for example,
perhaps something to do with the certainty of "sound money" and
affords in any case an interesting lesson about how business reacts in the
face of what many economists would describe as deflationary tendencies.
The ascent from 1897 may have something to do with gold or, more
1 The reader will reflect however that, in statistics as in life, the argument against Sin
is not, in good logic, weakened by the observation that the preacher sins himself.
2 The results of the two tests carried out are not shown in the graph. It should be men-
tioned, that United States and English clearings were among the series by the analysis of
which Mr. Kitchin first showed the existence of the cycle which we refer to by his name. He
also showed that "two or three" of them seem to form a higher unit (our Juglar).
556
BUSINESS CYCLES
plausibly, with the practices of trust companies, but would be under-
standable without either. Nothing indicates obviously the presence
of other than our evolutionary factors.
The long wave could not, in the absence of other information, be
discerned in this graph any more than in Chart XXIV, since our series covers
only the end of the second (1875 to 1897) and the beginning of the third
Kondratieff. But if we tried to fit, by the method of least squares,
LOANS AND DISCOUNTS
OUTSIDE N.Y.C.
1875 I860 1885 1890 1895 1900 1905 1910 1913
CHART XXV. — United States (see Appendix, p. 1000).
either a straight line or a second-degree parabola, we should discover,
as others have before us, that for some reason a "break in trend" occurred
in the nineties. If thereupon we proceeded to fit two straight lines,
the one to the interval from 1875 to 1897 and the other to the interval
from 1898 to 1913, we would find that the gradient of the former is
smaller than the gradient of the latter. It is submitted that, since we
know about the Kondratieffs from other evidence, historical as well as
statistical, and since that behavior of clearings conforms to what we
should expect it to be in those intervals, there is some point in inter-
preting it as a Kondratieff effect. This would not only solve the problem
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 557
of that break in trend, but also enable us to account for the fundamental
contours of United States clearings in that period by a single set of prin-
ciples. Whether this is more convincing than an explanation in terms of
gold production, is for the reader to decide.
The shorter cycles stand out well and show in the two intervals in
the way we should expect them to show in Kondratieff depressions,
revivals, and prosperities. The series is, in fact, extremely "responsive." *
We behold at the beginning the shrinkage of system expenditure incident
to the deep depression of the seventies. Then follow two very well-
marked Juglars (1879-1888, 1889-1897) within each of which three
Kitchins are clearly discernible. The strong swell of the prosperity of
the third Kondratieff dominates the picture after 1897 and tends to iron
out the other two cycles, which assert themselves nevertheless. The
relation of the clearings series to others follows from what has been said
before in this and preceding chapters. As regards timing, the reader
should again be warned not to form rash expectations about consistent
sequences or to draw inferences from a failure to find such sequences.
For instance, he may feel tempted, considering that system expenditure is
intimately related to the initiating impulse of the cyclical process and
that it is itself the immediate source of many of the symptoms of general
prosperity or depression, to expect that variations of clearings should
precede variations in most other cyclical quantities and, especially,
variations in price level. As a matter of fact, they mostly do, within
short -time fluctuations, by a few months. But even apart from internal
irregularities, the influence of speculative anticipation, the fact that
prices are made by contracts while clearings reflect actual payment, and
so on, there is no theoretical reason for expecting precedence of clearings
for all phases of a cycle but only for the positive ones, and for these only
with a proviso as to the simultaneous phases of the other cycles. Statis-
tical measurement of covariation by means of formal methods is bound
to give disappointing and inconclusive results,2 particularly if the method
used implies the single-cycle hypothesis.
Much more interesting than variations of total system expenditure,
are the variations of its two principal constituents — consumers' and
producers' expenditure. But even for postwar times we are far from
having adequate data from which to compile separate indices of them,
while for prewar times we must not be under any delusion as to the value
1 Of. Professor Crum's Interpretation of the Index of General Business Conditions,
Review of Economic Statistics, Supplement 2, September 1925.
2 The Harvard Committee's correlation between outside clearings and Bureau of Labor
Statistics prices at wholesale (both corrected for seasonal variation and trend), for 1903-
1918, is described as Fair with a four months' lead of poor consistency in the former
(Review of Economic Statistics, 1919, p. 184).
558
BUSINESS CYCLES
of any inference we may draw from such material as we have. Efforts in
that direction are nevertheless among the most urgent desiderata econo-
mists have to address to public and private organizations commanding
the necessary means. What we can do here falls very far short of what
might be accomplished even now.
A
f \
EXPENDITURE ON /
PRODUCERS* GOODS/
EXPENDITURE ON
CONSUMERS' GOODS,
V-
A
"'OUTSIDE
CLEARINGS
PRODUCERS' GOODS
EXPENDITURE
OUTSIDE CLEARINGS
A
\ :
V
CONSUMERS* GOODS
EXPENDITURE
OUTSIDE CLEAR NGS
1875 I860 1885 1890 1895 1900 1905 1910 1913
CHART XXVI.— United States (see Appendix, p. 1060).
A rough indication of the behavior of consumers' expenditure, the
most important figure of monetary theory, may be derived from the series
of total wages disbursed as represented by pay rolls, since this is not
only the biggest individual item in the sum total of incomes, but also
the one which may be expected to bear a closer relation to actual expendi-
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 559
ture on consumers' goods than any other. In spite of all qualifications
which have to be made but are too obvious to warrant restatement, we
may again compare wage bill and outside clearings (Chart XXIII).
The line at the bottom of the chart will help in interpreting the result.
The sharp rise of the clearings-pay roll ratio at the beginning of the
current Kondratieff is particularly instructive. Since the question of
how the sum total of taxable incomes is related to consumers' expenditure
is controversial, we will not press it into the same service.
Another indication of the behavior of the two main constituents of
system expenditure is given (Chart XXVI) by the index of physical
quantity of consumers' goods times the index of consumers' goods' prices
and the index of physical quantity of producers' goods times the index
of producers' goods' prices. Our chart is unsatisfactory in many respects,
and no confidence can be placed in the details of the picture. But how-
ever much progress in the direction of exact measurement we may expect
from more thoroughgoing investigation which government agencies or
research groups might undertake, it is extremely unlikely that it would
result in substantially different contours or that errors should have sys-
tematically worked so as to give spurious support to our theoretical
structure.
We notice, first, that Juglars and Kitchins show well in both the
dollar values of consumers' and in the dollar values of producers' goods,
which we assume to indicate consumers' and producers' expenditure.
In the case of the Kitchins it is, no doubt, necessary to recollect what has
on previous occasions been said about the likelihood of their being "ironed
out" in very strongly marked phases of the longer cycles, but they
remain always discernible, at least in rates. The rise of the third Kondra-
tieff, 1898 to 1907, is impressively reflected. Second, we observe the
difference in amplitudes of fluctuations, a familiar feature, which need
not detain us. Third, it is seen at a glance that producers' expenditure
dominates the contour of the clearing series. Fourth, producers' expendi-
ture can on the whole be said to precede, although not consistently.
For reasons repeatedly mentioned we would attach little importance to
this, even if our chart were as perfectly exact as it is defective. We have
seen reason to question the validity of causal interpretation of statistical
sequences; theoretically we cannot with any confidence expect that pro-
ducers' expenditure should precede in time as well as in logic, for, effects
on consumers' expenditure asserting themselves quickly, we might well
find roughly simultaneous covariation blurred by secondary and random
influences rather than a consistent sequence conforming to that logic;
and, finally, the case well exemplifies the possibility that a "logical"
sequence be inverted by anticipation, for even if consumers' expenditure
were the prime mover of the process, producers' anticipations might
560 BUSINESS CYCLES
easily make producers' expenditure move first.1 Without, however,
receding from this standpoint, we may still notice a possibly significant
fact. If the reader smooth both curves by freehand, he will get a picture
that roughly renders the Juglars and the Kondratieff effect. And
within this picture, precedence of producers' expenditure is unmistakable
and consistent. Inconsistencies of lag may thus be shown to be due
to the shorter fluctuations. Since these are particularly exposed to
disturbing influences and fundamental relations work ^ themselves out
more clearly in the longer cycles, that fact may mean something, after all.
Now, these four features certainly tend to justify — no proof is ever
possible by statistics alone — our decision to consider producers' expendi-
ture as the "active" element in total expenditure. We move on common
and familiar ground when we say that expenditure on plant and equip-
ment is the " active " factor in producers' expenditure. But if we go on to
say that expenditure on the creation of new production functions —
innovation — is the active element in total expenditure on plant and
equipment, we are leaving this terra firma, and there is every reason for
the reader to examine the nature of this third step. It is not, like the
two others, a formulation, but an interpretation of time-series fact. It
adds an explanatory hypothesis. But this hypothesis justifies itself,
by solving a problem presented by time-series fact — namely, the problem
of the wavelike character of expenditure on plant and equipment — and
by doing this, through its appeal to intermittent entrepreneurial impulses,
in a way that leads us out of the circle of perpetuum mobile theories.2
Moreover, it follows from a schema all the expectations from which
are borne out by time-series fact, where such fact is available. Finally,
it is fully verified by economic history: we simply know what innovations
were responsible for the humps in the curve of values of producers' goods
from 1879 to 1882, in the late eighties, and in the late nineties, or at the
beginning of the century.
Attention is called to the two lines at the bottom of Chart XXVI,
which present the fluctuations of consumers' and producers' expenditure
per dollar of system expenditure. As we might expect, they demonstrate
1 Compare again the discussion between Professors Frisch, Hansen, and Clark on
Capital Production and Consumer-taking, Journal of Political Economy, 1921-1922.
2 The problem may also be formulated as follows: How is it possible for prosperity to
start, let us say, from perfect equilibrium, without an impulse being first given to produc-
tive activity by consumer's expenditure? It would, indeed, be impossible but for inno-
vation's breaking through the existing system of production functions and creating an
opening for further outlay — new profitable opportunities for the expansion of producers'
expenditure, or, as we may call it, new "investment." Hence the mere fact of those
recurring bursts of "investment'* might be held to furnish verification of the presence of
what is, unless they are explained by external factors, their only adequate en use, the
element that induces all others.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 561
with particular emphasis what has just been explained. Both the rhythm
and the mechanism of our process — in particular, the character of pros-
perities as periods of investment and the character of recessions and
revivals as periods of consumers' harvests — assert themselves very
instructively. But there is also another movement — the long-time and
systematic change in the relation between consumers' and producers'
expenditure, which we already have had occasion to observe in the realm
of physical quantities, and which lends itself so readily to interpretation
in the sense of Boehm-Bawerk's theory. The fact that investment gains
on consumers' expenditure in prosperities without ever losing ground in
the long run, which stands out particularly for the period of the Kondra-
tieff upswing, accounts for the only long-run difference there is between
the behavior of producers' expenditure and clearings. Outside Clearings
in turn should now be compared with outside deposits and loans; but
before taking up this matter, we will digress, in order to discuss the
behavior of total income and of wages.
C. National Income and Wages. — 1. The familiar difficulties which
we all experience in defining National Income are of course due to the
fact that it is not a technical term wedded to one definite use but a word of
common parlance that is loosely used for a great many purposes which
cannot be served equally well by a single definition. For our purpose
we need not discuss the general merits or demerits of inclusion or exclusion
of annual services of durable goods or of the values of services rendered by
the individual household to itself or of the values of receipts in kind or of
the gains from cooperative purchasing, or any of the other questions that
arise in defining national income and in evaluating it statistically,
although some of them will have presently to be mentioned for the special
case of wages. At the moment, we will simply think of a monetary
quantity (hence, not identical with value of output) that consists of the
receipts of households from the sale of services, personal or other — sub-
stantially wages, rents, quasi-rents — plus profits and interest, but
exclusive of gains from the sale of capital assets, appreciation of inven-
tories, and so on. This total is equal to consumers' expenditure (on
durable as well as on transient goods), households' investments (for
definition see Chap. Ill, Sec. A), and tax payments, minus households'
net borrowings and expenditure of capital gains, plus sums that are not
spent at all (which we know can occur equally well in the cases of sums
earmarked for consumption, as in the cases of sums earmarked for
investment), and it makes, for most purposes, a highly inconvenient
composite. But it serves for our purpose.
The corresponding time series is obviously natural, primary, conse-
quential, systematic, and cyclical. It displays a result trend in the same
sense as system expenditure, that is to say, only because and as far as the
562 BUSINESS CYCLES
cyclical process of evolution expands the elastic strings of the monetary
ligamina for good. Barring this, there would be no result trend in it,
and national income in that sense would, in the process described by the
pure model, return in each neighborhood to its value in the preceding one,
which it is not superfluous to mention in view of an application to be
made of this proposition in the theory of the cyclical behavior of the wage
bill.1 In every four-phase cycle we shall still expect, with the usual
qualification about the simultaneous phases of the other cycles, that our
total increases in prosperity more than does the output of consumers'
goods. But expectation as to its behavior in recession is uncertain
beyond two points : we shall predict that if it increases at all, it will do so
at a rate that, on the one hand, is smaller than in the preceding prosperity
and, on the other hand, is smaller than the rate of increase in output of
consumers' goods. In depression it will, in general, fall — at least in
"deep" depression — and in recovery it will return to equilibrium amount.
The only data at all adequate for the purpose of compiling a series
of national income, are, for prewar times at least, those supplied by the
materials of the English and the Prussian income taxes.2 The first,
covering the whole period since Sir R. Peel's rein troduct ion of the
income tax, is the one we shall use, because the second, the figures of
which should, owing to the superior technique of Miquel's income tax act,
really be more valuable, covers only a little more than the years of the
third Kondratieff.3 Of course, the series of taxable income is not what
we want and can only in a very rough way indicate the movement of the
quantity we are interested in. Taxable income covers both less and more.
Since the wage bill or, more correctly, that part of the sum total of wages
1 Strictly speaking, it would, for that proposition to be borne out by statistical fact,
also be necessary that the sums saved and the sums paid in taxes from income were equal
in both neighborhoods. For instance, an increase of the income tax for the purpose of
increasing the salaries of public servants would increase national income as above defined
and an analogous proposition would hold true for savings. This is one of the reasons why
this definition has been called inconvenient in the text: a definition that entails the conse-
quence that national income falls if the sum total of net saving falls has certainly no claim
to general acceptance.
a There is, for the United States, an estimate by the Brookings Institution of national
income paid out, for 1901 and for 1910 to 1914; and by the Cleveland Trust Company for
1902 to 1909 (see Bulletin of the Cleveland Trust Company for Apr. 15, 1935). See also,
W. I. King's National Income and Its Purchasing Power, 1930. But in an extremely
rough approximation, railroad gross earnings may perhaps give an index; see A. H. Cole,
A Monthly Index of Railroad Earnings, 1866-1914, chart comparing this index with
Partington's quarterly series, on p. 41, Review of Economic Statistics for February 1936.
Professor Cole's index is reproduced on Chart XXIX.
8 There were other income taxes in German states, some of which — for instance, the
one of Saxony and the one of Baden — give useful indications. See a study on Einkommen-
schwankungen vor dem EJriege by the Institut ftir Konjunkturforschung, Vierteljahrshefte,
1927. Results do not differ from those that we are about to derive from the English series.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 563
that is not subject to the tax, could be added from 1860 on, and since that
part of "unearned" income that goes to income receivers below the
exemption limit may be estimated from the data of the Inland Revenue
reports, this is less serious than the effect of exemptions, allowances,
and abatements, on the one hand, and inclusion of various items that
ought not to figure, on the other.1 It must also be borne in mind, of
1780 1790
1800 1810 1820 1830 1840 1850 1860 1870 I860 1890 1900 191013
CHART XXVII. — United Kingdom (see Appendix, p. 1060).
course, that income from foreign sources cannot be expected to vary as
home-produced income.
The picture presented in Chart XXVII cannot hence be implicitly
trusted. It is, nevertheless, not very hazardous to say that it bears out
expectations as to Juglars in which total taxable income in fact displays
a tendency to rise in prosperities and to stay at about the figure reached,
or very little less, for the rest of each cycle. The reader can easily satisfy
himself of this by recalling the dates of the Juglar prosperities. Since
the technique of the assessment of business profits has a smoothing effect,
we shall understand that we see little of a Kitchin movement, but the
behavior of the series in the Kondratieff wave is contrary to expectation
up to 1873. 2 From that year on until 1898 we observe a remarkable, if
1 Much work has been done in order to improve the material and to distill from it
really relevant figures. See, in particular, Sir G. Paish's paper in the Journal of the Royal
Statistical Society, 1909; and Sir J. Stamp's British Incomes and Property.
2 The strong increase of total taxable income from 1860 to 1873 indicates the presence
564 BUSINESS CYCLES
rough, parallelism of "trend" with the production series, while there is
none with the series of production times price level. From about 1876
to 1913 parallelism with the wage bill is no less remarkable. This fact
is in itself sufficient to dispose of some of the more primitive errors about
the relation of wages and other incomes. The relation to provincial
clearings is what we should expect it to be.
The behavior of the series of total taxable income is obviously domi-
nated by business profits in the usual sense, as the reader may see by
referring to Chart XXVIII. The " profits" there charted are simply
from the data of Schedule D, which include not only interest, royalties,
and the like, but also a large amount of income that ought to be classed
with wages. The distinction made according to formal rules between
" earned" and " unearned" income does not, of course, help us at all.
We may look upon that item as roughly rendering the income from
"business." But it is far removed from profits in our sense, which form
an unknown and varying fraction of it. It is again by way of an inter-
pretation, which is not gleaned from, but added to, statistical fact, that
it can be asserted that entrepreneurial activity is what directly and
indirectly accounts for the waves we observe. For Germany we have
also a fairly long series of dividends (per cents of nominal capital), see
Chart XXXVIII. Dividends are of course not profits — still less are they
profits in our sense — but the relation is close enough to permit taking them
as an indication. Perhaps we need not stress the proviso usually made
about lag, for the state of affairs at the moment of the decision about
dividends in many cases influences that decision as much as the results
achieved in the preceding business year do. On the other hand, the
policy of equalizing returns to shareholders, which prevails in other
cases, naturally dampens fluctuations.
2. We must go more fully into the behavior of wages. There is a
considerable amount of information for earlier times, sometimes per-
mitting alignment into time series, practically always enabling us, in
spite of all difficulties about currency and prices, to discern the roughest
contour lines, at least for Germany and Western Europe.1 By them-
of a disturbance external to our model. Since the anomaly obviously links up with the
anomaly we observed in the price level series, we may attribute it to gold. When gold
production turned down, total taxable income assumed what we could conceive to be its
normal behavior.
1 E. J. Hamilton, Money Prices and Wages in Valencia, Aragon, and Navarre, 1859-
1500, 1986; and Prices and Wages at Paris under John Law's System, Quarterly Journal of
Economics for November 1986. See also, besides the familiar French and English works on
the subject, such as Thorold Rogers' or Levasseur's, E. W. Gilboy, Wages in Eighteenth
Century England, 1934; and Schmoller's handy survey (antiquated in part, of course),
Tatsachen der Lohnbewegung, in his Jahrbuch 1914, republished in the second edition of
his Grundriss, particularly valuable because of the insistence on the institutional setting,
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 565
selves, of course, those data mean very little, particularly with respect to
any welfare considerations. Recent work has done much to enlighten
us about the eighteenth century, notably in England, and more may be
hoped for in the near future. But at every step so far, we find room for
difference of opinion as to questions of fact.1 Mrs. Gilboy's book brings
out sharply that even England was then not yet, with respect to wages,
an economic domain in the theorist's sense and that wage rates at times
even moved in different directions in different parts of the country. We
do not venture to go beyond stating that, on the whole, money wage rates
started to rise, or in some cases continued to rise, after 1750; that they
rose during the prosperity phase of the first Kondratieff ; while real wage
rates on the whole rose very much less, if indeed they rose at all.2 For
the nineteenth century we rely on the work of Mr. Wood and of Professor
Bowley, which, for agriculture and building, also includes some eigh-
teenth-century data.
For this country we have scattered information for earlier times, most
of which is reproduced or noticed in V. Clark's work. Nation-wide
figures that are at all reliable do not date back of this century. We may
mention, however, Bulletin 499 of the United States Bureau of Labor
Statistics, which attempts to cover available material from colonial
times to 1928 and Bulletins 59, 65, and 77, for the time from 1890; the
National Industrial Conference Board studies on the subject, as well as
the work of Mr. Rubinow and Professors Hanson and Wolman; but
especially Professor Paul Douglas's treatise on Real Wages in the United
States, 1890-1926, to which the present writer is mainly indebted. We
are very badly off for Germany until almost the end of our period.
There are, however, more than a dozen individual series, of which the
wages of miners in the basin of the Ruhr, beginning in 1850, is the most
interesting one.3 We have confined ourselves mainly to the English
which goes far toward bringing out the true meaning of the wage data themselves. F.
Simiand's work, Le Salaire, 1'evolution sociale et la monnaie, essai de theorie experimentale
(sic) du salaire, 1932, much disfigured by methodological dissertations of doubtful value
and highly questionable theorizing, gives French wage rates per day from 1789 to 1930
and very useful documentation. We will mention, finally, the interesting article by D.
Knoop and G. P. Jones, Masons' Wages in Mediaeval England, in Economic History for
January 1933; see chart on p. 486.
1 See, for example, the controversy between Mr. Hammond and Professor Clapham
(Economic History Review^ 1930, and the latter's reply in the preface to the second edition
of his volume on the Early Railroad Age) about the state of things at the end of the eight-
eenth century and the beginning of the nineteenth.
2 Much bolder statements are made in Mr. R. Tucker's article in the Journal of the
American Statistical Association for March 1936. But see E. W. Gilboy's Cost of Living
and Real Wages in Eighteenth Century England, Review of Economic Statistics, 1936.
3 Two books may be mentioned- R. Kuczynski, Arbeitszeit und Arbeitslohn; and
Tyszka, Lohne und Lebenskosten in Westeuropa im 19 Jahrhundert. The very good
566 BUSINESS CYCLES
case. All data are primarily for money rates. Wage bill totals are not
available for most of our period and for most countries. But there is an
English series from 1860 on, which we owe to Professor Bowley.1 The
American figure we use is but an indicator of the behavior of the national
pay roll. Moreover, most wage data fail, even at their best, to represent
wages actually paid, which may differ quite considerably from official,
particularly trade-union figures, which sometimes veil a fall and at other
times represent minima only. It is believed, however, that most of the
conclusions of this section are outside of the danger zone created by those
and other defects in our data.
All sorts of difficulties arise previously to, and independently of, those
that are inherent to our material. We will mention two. There is, first,
a difficulty about the delimitation of the returns that should be looked
upon as wages. This term means different things to the economist and to
the sociologist, and the economic class of wage earners is much wider than
the social class to which we refer by the same term. But although part
of the salaries and other emoluments of managers and executives should,
by the economist, be included in wages, another part is a rough contractual
equivalent for, or share in, profits in our sense. An analogous difficulty
arises about the incomes of independent business men, while professional
incomes are almost exclusively wages. In practice, of course, we simply
use the material we have. Second, excepting series of wages paid to
homogeneous groups of workers in a given firm or industry or neighbor-
hood, the only natural series we have is the series of the money wage bill.
All others are synthetic ones and raise, anterior to questions of deflating,
index problems incident to the concept of the average wages of a working
population.2
We distinguish the Sum Total of Wages (or Payroll or Wage Bill),
Wage Rates (per unit of time or product or per head of population
employed or of population seeking employment, ideally, per man-hour)
Swedish data are, again, in a class by themselves; see G. Bagge, Wages in Sweden 1830-
1980, vol. II of the Stockholm Economic Studies.
1 Tests of National Progress, Economic Journal for September 1904, continued by Pro-
fessor Pigou, see Table III of the appendix to Industrial Fluctuations. Professor Bowley,
in kindly permitting the writer to use those figures, expressed a wish that it should be stated
that he has revised his treatment in a book which has since been published by the Cam-
bridge University Press.
2 See particularly A. L. Bowley, Notes on Index Numbers, Economic Journal for June
1928, Sec. 6. Differences in behavior of different classes of wages, both in the short and
in the long run, which sometimes amount to differences in direction, are, on principle, still
more serious than the regional differences, which in some cases are also relevant to the
cyclical process (migration of an industry may be an important innovation; migration of
laborers from country to towns and vice versa is in part a cyclical phenomenon). Geo-
graphically, real wages differ very much less than money wages. The item of rent is some-
times sufficient to produce approximate equality in the former over wide areas.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 567
Labor's Share in the national income or in the value of a unit of product.
Wage Bills and Rates must be considered both in money and in real
terms. For this purpose, they have to be deflated by an appropriate
index of the cost of living, an operation which can, at best, yield but an
approximation to the actual command over commodities enjoyed by
the various classes of workmen whose wages enter into the average.
There is, however, also meaning to another concept of real wages —
which should be called by another name, say, Corrected Wages — viz.,
money bill or rates divided by an index of wholesale prices. This
relation has, in itself, an obvious cyclical significance. But inasmuch
as that index represents the variations of the price level, it acquires an
additional meaning — that of an indicator of the course of wages from
which the influence of the monetary parameter has been eliminated.
It should be clear that a cost-of-living index, though in most cases it
will give broadly similar results, is not in principle qualified to render
that service on account both of the commodities that enter into it and
of their weights.
The wage bill is not the whole of the income of labor, which consists
also of other items both in money and in kind. But for our problem
they do not call for consideration in spite of the fact that the importance
of public expenditure for purposes directly increasing the monetary
and real income of the working class increased considerably in the last
decades of the epoch under survey, though not to the extent that it has in
postwar times. We similarly neglect the fact that wages do not con-
stitute the whole cost to firms of employing labor. Also, we assume that
the merely statistical effect on the wage bill of small shopkeepers, artisans,
and so on, "migrating" in and out of employment without changing the
economic nature of their income, is negligible. A measure of variation
of real wage bill should be made to include increase in voluntary leisure,
hence, be corrected for reduction of hours of work, provided, strictly
speaking, that hours are not reduced in order to distribute the burden
of unemployment. Our failure to do so involves understatement of
the historic rise in real wages. Average wage per employed workman
corrected for unemployment can, as long as the population seeking
employment increases, be taken to indicate how the wage bill has at-
least increased or at most diminished.
3. All wage series are systematic and cyclical and display a descriptive
trend, which is the distorted picture of a result trend. What they
describe is a primary and consequential phenomenon that is causal
to some of the secondary processes. The facts for which we have to
account are easily read off from Charts XXVIII, XXIX, and XXX.
It is seen at a glance that on the whole they bear out the expectations
which our discussion of the working of our model would lead us to form.
568
BUSINESS CYCLES
The two important exceptions should be disposed of at once. First,
we should have expected money wages — both bills and rates — not
indeed to fall, but to increase at a smaller rate from 1856 to 1873 than
during the preceding Kondratieff prosperity. We find in fact a setback
in the late fifties, but a strong rise from 1862 on. In America (see, also,
Professor Mitchell's weighted index of daily wages; but it is safe to infer
that national pay roll rose in at least the same proportion) this is of
I860 1865 1870 1875 1880 1885 1890 1895 1900 1905
CHART XXVIII.— United Kingdom (see Appendix, p. 1061).
1910 I9B
course due to an "external factor," the Civil War inflation. We have
discussed the point in our historical survey and will merely emphasize
again that the fall in the seventies did not fully correct that deviation.
Wages remained permanently on a new level, explainable by the monetary
event which also influenced subsequent cyclical behavior. As far as
this goes, there is close analogy with the aftermath of the World War.
But we find the same phenomenon, though less marked, in England
(and the rest of Europe), where it must be attributed to gold. It is
significant — both for the theory and the policy of wages — that real rates
and bills do not display a corresponding departure from expectation.
Second, while in this country and, as far as it is possible to judge, in
Germany, money bills and rates behaved according to expectation during
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 569
the prosperity phase of the third Kondratieff, in England they rose
rather less than we should have expected. This tallies with other
peculiarities of the English situation of that time but is less easy to explain
than the other case. It coincides, however, as has been pointed out
before, with greatly increased public expenditure and taxation. Other-
1850
1860 1870 1880 1890 1900
CHART XXIX. — United States (see Appendix, p. 1062).
1910
wise, the cycles are as we should expect. The Kitchins are not more
than just recognizable, the Juglars very well marked, the incomplete
Kondratieff s show in the same way as in clearings. It is particularly
interesting to observe the rhythm in real rates corrected for unemploy-
ment, and the way in which Juglars assert themselves, in the various
Kondratieff phases, in the deviation from the nine-year moving average;
see Chart XXXI.
Bearing in mind, as we must throughout, that we are dealing with
yearly and, moreover, not very reliable figures, we observe that absolute
fall in the British monetary wage bill (Chart XXVIII) occurred, in the
two most serious cases, during those years of the depression phase of
570
BUSINESS CYCLES
the second Kondratieff in which its sweep was not interrupted by Juglar
prosperities. The figures for 1874 and 1875 display but moderate
decrease from the peak of 1873 (485 million pounds; 1874, 470; 1875,
465) but the fall went on through the Juglar recovery, right to the
threshold of the prosperity of the next, although it at no time brought
1850
I860
1870
1880
1890
1900
1910 1914
CHART XXX. — Germany (see Appendix, p. 1062).
the wage bill back to the figure of the previous neighborhood of equilib-
rium (365 million pounds). This prolonged fall is not surprising, how-
ever—nor is the fact that the next peak (1882-1883, 470 million pounds)
is lower than the one of 1873 — and perfectly accords with our cyclical
schema. Our first impression to the contrary vanishes as soon as proper
account is taken of the principle of interference, when the phenomenon
is readily recognized to be quite "regular." This is not to say that gold
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 571
production, either directly or by virtue of its responsibility for some
of the excesses of 1872, had no share in the matter, but only that it
cannot have done more than accentuate what would have happened
without its decrease and what might have been toned down had it con-
tinued to increase or failed to increase previously. The figures for 1884
and 1885 (the fall was arrested in 1886, and the 1887 wage bill increased
again) reflect the conditions that we may, according to our schema,
expect to prevail in the depression of the Juglar at the bottom of a
Kondratieff . But there are within the series four other cases of absolute
1850
I860 1870 1880 1890
CHART XXXI.— England (see Appendix, p. 1062).
fall in wage bill. The years 1868 and 189& do not call for comment.
We also readily understand that a fall may occur in the course of a
Kondratieff prosperity, as a result of a "crisis," provided it be as short-
lived as it was after 1907. But there is a prolonged fall in the English
figures for 1901 to 1904, which is contrary to expectation and to which
nothing corresponds in this country except a small setback in 1904.
It is of course the same phenomenon that was noticed above, but with a
new trait added. Most years of unusual increase are easily identified
as years of Juglar prosperities, though there are some among them that
belong to either revivals or recessions.
We do not expect, nor do we find, any significant lag in monetary
wage bill behind comparable aggregates. On the upgrade it rather
precedes, if anything, the index of wholesale prices. There are cases,
such as the behavior of the British series in the nineties, that illustrate
572 BUSINESS CYCLES
almost ideally, and by so doing testify for, our view of the mechanism
at work. In that particular case, wage bill recovers from 1893 while
wholesale prices are still falling and breaks into a stronger rate of increase
in the same year in which wholesale prices turn upward. In fact, as the
reader will recall from our theoretical discussion, the factors that bring
about recovery and the different factors that bring about prosperity
may indeed make themselves felt at an earlier date in some of those
elements of the system that move in response to speculative anticipation,
but their mechanical effect must show early and unmistakably in the
wage bill.
4. If there were no unemployment in the neighborhood of equilibrium
from which a cycle starts, and if population were constant, the same
would be true of money wage rates. Since these conditions are not ful-
filled, we shall expect a lag in rates on the upswing, which measures the
amount of preexisting unemployment. We find it. In 1868 wage
bill increases while rates remain constant. In 1880 wage bill goes up
while rates fall in that and even the following year. The year 1887
presents a similar picture. But we find, although annual figures are
inadequate for the purpose, rather less of lag than we might expect.1
The English series covers six Juglar prosperities, of which four set in
toward the end of the year, so that an annual average may well produce
a spurious lag, as well as veil a real one. In the other two cases no lag is
observable, and no lag greater than a few months can have been present.
The effect shows rather in that wage bill rises, in the beginning of an
upswing, more strongly than wage rates and that later on these rise
oftener than wage bill at an increasing rate. It is worth mentioning that
that lag, such as it is, which has been so often emphasized and even
overstated, docs not present any new problem for us and docs not require
introduction of any cause extraneous to our mechanism. But it is still
more worth noticing that the same holds true for the "lag" in wage
rates on the downgrade, which is more pronounced. Both bill and rate
increase in 1866, 1873, and 1883; rates stay up in 1884 when wage bill
falls, and similarly in 1893 and 1900. Wage rate does not reflect 1907,
although wage bill does; neither reflects the recession of 1913, in which
year rates rise.
Now most economists are inclined to call upon such factors as sticki-
ness in order to explain this. Without denying the validity of such
explanations for many cases — although those factors cannot have had,
at least up to the turn of the century, anything like the importance
they were to acquire later on — we must still remember that, while there
1 The lag is, however, clearer in some particular industries. Compare, for example,
Hooker, Relation between Wages and Number Employed in the Coal Mining Industry,
Journal of the Royal Statistical Society, 1894,
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 573
is strong reason to expect a fall in bills and rates during "deep" depres-
sion, there is much less reason for expecting that they should fall in
recession. The opinion that wages should fall because, and to the same
extent as, prices fall and that their failure to do so is necessarily an
additional source of disturbance, is incorrect. If reality conformed
to our model, as designed for the purposes of the First Approximation, the
monetary wage bill would indeed have to return to its figure in the
preceding neighborhood of equilibrium, and rates would consequently
have to fall, unless vicarious unemployment is to ensue, at least pro-
portionately, but more than that if the population seeking employment
had meantime increased. The fruits of innovation and induced develop-
ments would entirely be harvested in the form of increased "purchasing
power" of the unit of income. But since, as we have seen, autodeflation
never does full work, and since as far as it does work, the requirements
of the conquest of new economic space counteract its effects, there
need riot be a fall in bill or rates at all: we may find instead a continued,
though weaker, increase. Which we are to expect in any particular
case, is, for a normal recession and before depression casts its shadow,
indicated by the behavior of the deposit figure. If this does not fall,
wage bill and, barring increase in population, rates need not. If it
increases, they can and will1 also increase, without causing disturbance
as they would if failure to fall were due to stickiness or increase were
due to, say, an act of legislation. Wages, bills or rates, need not even
necessarily fall during the depression phase, although they are much
more likely to. They will not, of course, if the depression of a cycle
coincides with a sufficiently strong prosperity or revival of another,
but it is more important to note that they may stay up, even barring
such coincidence, if panics and spirals do not play too great a role.
In the only Koiidratieff depression which we have before us, the English
wage bill actually increased by about 25 per cent, and no explanation
external to the mechanism of evolution is needed to account for this.
1 This is, of course, not equivalent to saying that if the deposit figure be made to
increase, by central bank action or by any other influence external to our process, wage bill
will increase also. But we shall understand that so pronounced a covariation — see, for
instance, the covariation of Pay Rolls with Net Outside Deposits minus investments in
this country, Chart XXIII — should have invited causal interpretation. The fact, how-
ever, that covariation is particularly good if deposits are cleared of the influence of the
item most completely under the control of banks — investments — clearly tells against it.
Perhaps it is not useless to state once more in a perfectly general form the problem involved.
We observe, let us say, any time-series material consisting of the time variables Xi • • • xn.
We find that one of them x% stands in an invariant relation to another, x,. It is submitted
that in the absence of any knowledge about the relations subsisting between all the variables,
xi ' • • xnt we are not justified in inferring from such a finding that any arbitrary change in
Xi will be accompanied by the change in xjt which actually does accompany it within the
system observed.
574 BUSINESS CYCLES
What autonomous monetary influences there were, worked the other
way.
Chart XXIX presents evidence that, as far as it goes and as far as
it can be trusted, justifies similar conclusions for America. Railroad
Earnings and Failures (liabilities) have been inserted to provide links
with the general business situation. The closeness of the covariation of
the former, as well as of "dividends" (see Appendix) with wage bill,
needs no emphasis : the dominating cyclical factor in wage bill is of course
employment. Behavior of corrected wage rates is according to expecta-
tion, but the strong increase in corrected wage bill, 1895 to 1908, is not,
and must be explained by peculiarities of the American environment,
immigration among them.
5. A question, however, suggests itself: Has not the above analysis
gratuitously assumed that the cyclical process leaves the relative marginal
significance of labor in the productive process unchanged? For only
in this case it follows — to return, for the sake of simplicity, to the pure
model, also assuming population to be constant — that money wage
bill and wage rate will be at the same figures in the new equilibrium,
at which they were in the old. Since most innovations are not only
labor-saving themselves1 but induce adaptations that are also labor-
saving, there seems to be no reason whatsoever to expect this; hence,
the wage bill may decrease whatever happens to the total income. We
therefore meet again the problem of machinery and labor — it is con-
venient to confine ourselves to this special case — which we have already
met twice: in our discussion of unemployment and in our discussion on
saving in an otherwise stationary process. But we meet it now in its
fundamental aspect, for technological unemployment is but a special
and, moreover, as we have seen, a temporary form of the effect of tech-
nological improvement on the wage bill. We meet it, also, in the new
setting provided by the process of evolution. We cannot any longer
reason on invariant production functions; nor can we any longer deal
with the isolated effects of laborsaving innovations. We must take
account of all the elements of our process and, in particular, of the
emergence of new investment opportunities not due, as it was in the
case of saving, to mere fall in the rates of interest. The usual arguments
both for and against the necessity, likelihood, or possibility that total
monetary wage bill will be increased or decreased by technological
improvement, thus become as irrelevant as, within their proper sphere,
they are inconclusive.
1 On the concept of laborsaving devices, see Mr. Hicks' Theory of Wages, p. 121.
Accepting his definition, we call an innovation laborsaving if it increases marginal product
of "capital" more than that of labor. It need not actually diminish the latter, and cor-
rected wage bill need not fall absolutely.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 575
But all the more conclusive are the facts. Although the wage bill
and the share of total income it represents fluctuate in the course of
cycles (even if we exclude from total income, as we should, gains and
losses from the sale of capital assets, in particular from speculation in
land and securities), its long-run relation to total income and to most
other monetary aggregates seems substantially stable (see Charts
XXVII and XXIII). We may infer that an absolute amount of wage
bill would, as between neighborhoods of equilibrium, be constant,
if deposits, clearings, and total income were. This of course implies
invariance in the long run, of the money wage bill, both to innovation
as such and to the relative increase, within total output, of equipment
goods it induced. Now our data are not too good to make it impossible
to attack the fact. If, however, we accept the evidence, such as it is,
then the next possibility is to say simply that it is very remarkable that
a relation which theoretically could fluctuate anyhow yet remains
constant, but that it is so. And this the reader had better do, unless he
accept the principle of the author's theory of interest, which would
offer a theoretical explanation: if it be true that in an ideally perfect
competitive equilibrium no assemblage of physical goods yields any
net return, then it follows that there must be, on the system's way toward
such a state, even if it be never reached, a tendency of all incomes,
except payments for the services of labor and natural agents, to vanish.
The "wage and rent bill" would tend to absorb, with a lag, the whole of
national income, and any failure of wage bill to keep, in successive
neighborhoods of equilibrium, a constant proportion to it could only
be due, either to change in the relation of marginal productivity of labor
to that of natural agents, or to varying degrees of imperfection in the
neighborhoods for which the figures of wage bill are being compared,
neither of which would in general be of an importance sufficiently great
to dominate the findings.1 Without appeal to this theory, it is still
possible to argue that, "real capital" being itself a product, its price
will tend to be a minimum whenever the system is approaching equilib-
rium. But if that minimum contains an element of interest, no general
proposition is possible about absolute or relative shares in the national
income, although it remains true that no such minimum principle operates
in the case of wages.
1 Professor Leontief objected to the above argument on the ground that, since the
author does not hold that interest or return to real capital — net marginal value produc-
tivity— is actually zero in the neighborhoods that we observe in real life, the result does not
follow, even if his theory be accepted. This is not so, however. If interest fails to become
zero because it represents a payment that is a necessary element of the productive process,
a payment for a distinct productive service for instance, then it becomes an open question
what happens to the relative importance of that service. If interest fails to become zero
for other, for instance frictional, reasons, then there is no room for such an argument.
576 BUSINESS CYCLES
The fact, whether explained or a mystery, that wage bill behaves as
if it were a total monetary aggregate, such as the sum of household
incomes, while we know it to be but a variable part of it, together with
what has previously been said about price levels, suffices to account for
the cyclical behavior of real and corrected wage bill and, with the obvious
qualifications, of real and corrected wage rates. In particular, we saw
how these quantities fared in recession and depression. That behavior
bears witness to the old proposition so familiar to the common man,
that the working class is, as a whole and on the whole, better off in
times of falling than of rising prices. This needs to be qualified, first,
for the fall of prices in "deep" depressions; second, for the rise of prices,
if any, in revivals, and third, for the rise of prices in the upswing of shorter
cycles that lie in the downgrade of the longer ones. Nevertheless, it
expresses a broad truth. To be sure, this truth is usually made to rest
on mere stickiness and the simple arithmetical effect of falling prices on a
given income. This is sufficient only for other than the cyclical cases —
cases of gold or government inflation, for instance — which are historically
associated with misery. But our argument shows that there is some
foundation for it also with reference to cyclical variations in price level.
As in similar cases, the true contours can be expected to reveal them-
selves most clearly in the Kondratieff. So they do. Information is
available for England, that would not warrant positive assertion, yet
suggests that this holds beyond the span covered by our series. Real
wage bill, as well as rates, very probably rose somewhat from 1775 to
1815, that is to say, in the later stages of a Kondratieff revival, the whole
of a Kondratieff prosperity, and in a little more than a Kondratieff
recession that was distorted by government inflation and hence cannot
be accepted as normal. Thereafter there was, until the forties, little
change in incomes, while cost of living fell so considerably that it is not
likely that better information about unemployment would invalidate
the inference that real wage bill, as well as rates, rose much more than
they had risen before. In the forties real rates cannot have increased
considerably. During the period of rising prices that followed, com-
pleting the Kondratieff upgrade, but outlasting it for reasons we know,
corrected rates certainly, and real rates possibly, fell (see Chart XXVIII)
until the top of the grade was reached — though, because of increased
employment, the real wage bill almost certainly rose — and then, from
the top (1856) to 1897, took the stride — nearly 80 per cent — which
created the modern standard of life. There was hardly any increase
in real rates in the period 1898 to 1913 taken as a whole, though of course
there was in real wage bill. Normally, and with but a moderate number
of exceptions, real bill and rates increase in season and out of season. But
it is the increase in the three last phases of each of our two Kondratieffs
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 577
that is mainly responsible for the observed trend in real rates. Since
this trend is the effect, however distorted by external factors, of the
working of our mechanism, it may also be identified as a result
trend.
From the above analysis it may be inferred that the sequence of
cyclical situations tends to produce wage rates and wage bills which are,
with due qualifications as to the many cases of indeterminateness that
arise, adapted to them in the sense that they do not in turn require or
induce a distinct process of adaptation on the part of other elements of
the system. In particular, it follows that occurrence of the upper turning
point or of depression is not normally such an adaptation to preceding
wage rates, enforced, for example, by any failure of the "purchasing
power" of laborers to develop in a manner that would enable them to
buy the increment of product resulting from every step in the evolu-
tionary process. Nor is it true that money wage rates or bills would
have first to rise, or to be increased by public authority or by the pressure
of organized labor before the system can move up from the lower turning
point. On the other hand, we have also seen that the mere fact of wages
rising in prosperity to the extent appropriate to the upward shift of
"demand curves for labor" is not the "cause" that turns prosperity into
recession, and that the failure of money wage rates to fall during the
latter phase has normally as little, if anything, to do with pushing the
system into depression as such fall as occurs in depression has to do with
helping the business organism on to the path of recovery.
These statements, however, do not imply anything about the effects
of those levels and variations of money wage rates that are not, as we
have put it, produced by the working of our process but are imposed on
the system. Such imposed rates may — although they need not; there
are plenty of reasons why, in a given instance, the system should fail to
give effect to its tendency to produce the adapted rate— spell disturbance,
i.e., enforce additional adaptations of other elements of the system.
But even if they do, this in itself only means that they disturb a process
which by virtue of its very nature creates, although it also eliminates,
disequilibria. Interference with such a process need not intensify these
disequilibria but, on the contrary, may mitigate them. We have not
even, by denying that the cyclical behavior of wages, such as our process
tends to produce, is causal to the occurrence of recessions, depressions,
and revivals, completely prejudged the question whether the first two
could not be toned down and the third facilitated by the imposition of
appropriate "inadapted" wage rates. No particular dignity is claimed
for the adapted rates. Our analysis refutes, indeed, some of the argu-
ments that are commonly adduced in favor of high or low wage policies.
But it does not preclude others. Satisfactory treatment of this problem
578 BUSINESS CYCLES
requires, however, the whole apparatus of general theory and cannot
be attempted here. And only a few aspects of it will be presented in our
discussion of the postwar epoch.
D. Deposits and Loans. — Returning to the argument of Sec. B, we
will now discuss the behavior of the sources of system expenditure, i.e.,
of (the cash and) the balances or "deposits" of firms and households,
and thereby also gather up the threads of the analysis of Chap. Ill,
Sec. D, as developed in Chap. IV. Perhaps it is not superfluous, speaking
of " sources" of expenditure, to advert to the metaphorical character of
that phrase and to the misleading associations to which it may give rise.
We are not now moving toward the origin of the cyclical process but, on
the contrary, away from it. It is precisely in order to avoid any implica-
tions about mechanical effects being exerted on the pulse of business by
the "flow of funds" that the subject of balances has not been taken up
before the subject of the expenditure they finance. For the same reason,
discussion of the behavior of banks and of the role of central banks will
be deferred to a still later stage. In this section we are going to look
at financing from the standpoint of firms and households only.
1. Any act of expenditure may be financed:
a. By Previous Receipts. The reader will recall the proposition that
in a process which merely reproduces itself at constant rates (stationary
process) all acts of expenditure could and would be so financed. Now
the bulk of a nation's business, at any given point of time, consists of
transactions that repeat with adaptive variations the transactions of
previous periods of account. For some purposes it is convenient to split
each of these transactions into an exact repetition and a plus or minus
variation, and to look upon all these exact repetitions, on the one hand,
and upon all the corresponding variations, on the other hand, as separate
classes of transactions to be distinguished from a third class, which con-
sists of new transactions — mainly, those effected by entrepreneurs or
induced by entrepreneurial activity. At the level of abstraction appro-
priate to our pure model we applied that proposition to the first class
of transactions, the repetitions. But we cannot do so any longer. We
have, indeed, to recognize the fact, much more frequently observable in
Continental Europe than in England or the United States, that many old
firms actually run their business on "owned" cash or "owned" deposits;
but this is now, even for exact repetitions, merely the limiting and not
the general case: most firms, old as well as new, currently borrow and
repay, even within the most ordinary business routine.
This practice is, as we have seen (Chap. Ill), a by-product of the
process of evolution through which borrowing and credit creation intrudes
into the old strata of the economic system, partly because balances
created for the purpose of financing innovation are hardly ever completely
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 579
eliminated by autodeflation, partly because other balances are created
for the purpose of financing the expansion of the noninnovating sectors
which comes about in response to the impulse given by innovation. What
we have called the Conquest of New Economic Space thus expands
permanently, though not in proportion to physical output, the monetary
system and at the same time creates that gap between producers' expendi-
ture and receipts which it then becomes the most "regular" business of
banks to bridge. A little reflection will show that this gap, which is so
familiar a phenomenon as to seem hardly to stand in need of explanation,
is in fact entirely due to "progress" and would not, except as a result of
previous waves of progress, be present in a stationary society. But,
although not extraneous to our model, the fact that, in a society the
monetary process of which has become adapted to evolution, repetitions
may require financing as much as plus-variations or new transactions
affects our expectation as to the behavior of balances. So far as repeti-
tions are actually financed by owned cash or original deposits, this cash
or these deposits do not, in consequence of a decision to restrict opera-
tions, disappear but simply become idle. But so far as they are financed
by borrowing, the corresponding balances do not become idle but dis-
appear in that case through repayment or else fail to emerge altogether.
The businessman's decision and its economic results are the same,
whether it is "quantity" or "velocity" of deposits that is affected, but
the statistical picture is different.
While, however, we have to recognize that mere repetitions also
present a distinct problem of financing, the problem of financing new
transactions may frequently be solved by previous receipts being deflected
from the channel into which they have previously been directed. A
man, who has been in the habit of setting aside a yearly sum in order to
buy a new motorcar every five years, may at the end of one of these
five-year periods decide to buy an aeroplane instead. Not only system
but also consumers' expenditure or, had we chosen an example from the
sphere of production, producers' expenditure remains unaffected in case
of such Deflection.
6. By Overspending. This term is to mean allowing one's balance to
fall below the amount appropriate to the requirements in the previous
neighborhood of equilibrium. This will affect system expenditure but
not total balances. The opposite case, Underspending or, as we have
previously called it, Nonspending, occurs if cash balances rise above
that amount and is usually referred to as "Hoarding." This word,
however, lends a wrong color to an extremely simple state of things and
should never be used except in order to designate the thing it really
means — still observable in India, for instance — which, sociologically and
economically, is perfectly definite and presupposes a characteristic
580 BUSINESS CYCLES
attitude toward money entirely foreign1 to capitalist society. We also
recall that it is misleading, although for other reasons, to express the fact
of underspending by reference to any demand for balances, which is in
this case purely fictitious, or to call it saving. Release from under-
spending is, of course, an important method of financing the expansion of
recovery. It should be observed that, although, given the sum total of
cash items of households and firms, all of them cannot increase or decrease
at the same time, yet the inactive part of them can.
c. By Selling Assets. If an asset be sold to a bank (to this we refer
as "member bank investment")* customers' balances will be increased
unless the bank neutralizes the effect. And if it be done in order to
finance an act of expenditure, it will of course lead to an increase in
system expenditure, although it would not by itself and in the absence
of such intention have that effect. The answer to the question by how
much it will increase it depends on the behavior of those firms and
households that are the recipients of the sum thus procured — hence
on the cyclical phases into which that event happens to fall. If an asset
be sold to another firm or household, it will not increase the sum of
balances but may still increase system expenditure if it is the occasion
of turning an inactive deposit into an active one. If a bank sells to
firms or households, the balances of the public are decreased.
d. By " Temporary Investment" This is the term given in Chap. Ill,
Sec. A, to signify using, or borrowing from another firm or household, a
sum which is and remains intended for another act of expenditure lying
sufficiently far off to make it possible to replace or repay that sum before it
will be wanted. We assume in sufficient approximation to fact that Tem-
porary Investment acts exclusively through the open market or the stock
exchange. It increases system expenditure but not the total of balances.
e. By Issuing Promises to Pay. These may be issued, for example,
in the form of bills of exchange, provided they are actually accepted in
payment and circulate (see Chap. Ill, Sec. D, 3). This item we shall
neglect, however.
/. By mining or importing the monetary metal and having it coined, or
exchanging it for a deposit, or, what economically comes to the same thing
(as has been pointed out by Wicksell), forging money. This increases
deposits and the bank's cash by the same amount. If not deposited,
the money may only increase the outside circulation.
g. By Borrowing from a Bank. This, in England and the United
States, increases deposits immediately (with a qualification as to over-
1 The only cases in capitalist society at all akin to true hoarding are those in which
lawful money is held from a distrust of banks, or some kind of lawful money from distrust
of other kinds, or some commodity from distrust of all kinds. But these phenomena are in
a class by themselves and entirely different from nonspending.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 581
drafts), and in Germany as soon as payments to firms or households are
effected.
h. By using one's own, or borrowing some other household's or firm's
uninvested savings or accumulations. By borrowing we also mean to
include methods which do not technically come within that term, such
as the formation of partnerships.
1. By Taxation and the Issue of Government Fiat. As a method of
financing business these items will now be neglected, although we have
repeatedly met them in this role in the historical chapters.
j. By Buying on Credit. But we will in general assume, although
this is often far from the truth, that in this case sellers borrow from banks
equivalent amounts. The same applies to the opposite case, in which
the burden of financing production is shifted by payment being exacted
in advance of delivery.
k. By Borrowing from Abroad. This acts exactly like / if it leads to
imports of gold, and in exactly the opposite way if the proceeds of the
loan buy foreign commodities.
2. Since all these methods of financing combine to shape the behavior
of balances in the phases of the cyclical process, and since there are so
many external factors that are bound to assert themselves, it becomes
doubtful whether the contours to be expected from our model will show
at all in the series of deposits and of bank loans. These expectations
need not be restated. Firms' and household's (cash plus) balances are
primary and consequential and make a natural, systematic, and cyclical
series, which should behave like system expenditure and display a result
trend in the same sense. The same holds for member banks' loans and
discounts. Modifications are necessary even in the case of the pure
model and will be noticed presently. But it is interesting to see how far
reality conforms to what seems to be a theoretical construction very
remote from actual fact. Chart XXV supplies the answer. The
covariation of deposits, loans, clearings, and of all three of them with pig-
iron consumption and the production of equipment, though far from
perfect, is striking. The reader is invited to satisfy himself that it could
hardly be expected to be stronger, the material being what it is, if there
were no factors or mechanisms at work other than those embodied in
our model. Our problem is not so much how to explain deviations from
that expectation as how to explain a conformity so much beyond reason-
able hope.
For both these purposes it is necessary to bear in mind the short-
comings of our material, which make it impossible to place much con-
fidence in anything except the roughest contour lines. Our analysis is
based mainly on American data, neither English nor German statistics
being at all adequate for our purpose, although they may be used as
582 BUSINESS CYCLES
supporting evidence and although more valuable information is available
for Germany since 1901. * Those American data refer to national banks2
and thus represent a sample of varying importance, which cannot be
trusted to reflect the condition of all commercial banks.3 The series of
individual deposits less clearing-house exchanges, which is the one that
has been primarily used, substantially gives the balances of firms and
households that we want, since it excludes government and interbank
deposits — although deposits due to savings banks were included by
some national banks until Apr. 26, 1900 — but the device of using individ-
ual deposits of national banks outside New York in order to reduce
the influence of speculative and other transactions which we wish to
exclude is of course as unsatisfactory in this case as it was in the case of
clearings. The fact that short fluctuations in deposits and in loans and
discounts outside New York City are inversely related to those in New
York deposits and loans and discounts tends, however, not only to
reassure us on that point, but also to show that the bankers' bank func-
tions, which we attribute to the national banks of New York, was not
completely overshadowed by their commercial and industrial business.
In comparing this series of outside deposits with our clearing series,
which represents a different sample, we are, of course, opening up another
source of error.
1 See Professor A. Hahn's analysis in Vierteljahrshefte zur Konjunkturforschung, 1927.
German bookkeeping methods and German terminology increase the difficulty of inter-
preting German banking statistics, which for the nineteenth century are also unsatisfactory
for other reasons. What the writer believes to be acceptable indicators of the behavior of
balances plus hand-to-hand circulation are, however, plotted on the pulse charts.
2 National bank data have been analyzed by many students. There is however one
outstanding contribution, to which the writer wishes to acknowledge obligation and to
refer the reader for many points, regional differences among them, that cannot be dealt
with in the text: A. A. Young, An Analysis of Bank Statistics for the United States, reprint
of articles from the Review of Economic Statistics, 1928, especially pp. 1-7, 21-32. Also
see J. P. Norton, Studies in the New York Money Market, 1907, and O. M. Sprague, Crises
under the National Banking System. There are, of course, figures for other banks, both
since and anterior to the establishment of the National Banking System. Some of the
older material has been generously put at the writer's disposal by Professor Cole. Refer-
ence should be made, also, to W. B. Smith and A. H. Cole, op. cit. Estimates of total,
time, and demand deposits, 1834-1935, have been made by Dr. C. E. Thomas; they will be
presented in the writer's book on money. For the purpose in hand it has seemed best to
keep to comparatively safe ground. No indications have been encountered to warrant
a suspicion that results are affected thereby.
8 The growing importance of trust companies in the nineties and particularly after
1899, which, with the rise of the Kondratieff, "began to invade, on an extensive scale, the
field of deposit banking" (A. D. Noyes, Forty Years of American Finance, 1909, p. 367)
and increased their deposits from $198,000,000 in 1898 to $834,000,000 in 1906 (ibid.,
p. 368), constitutes a further and very serious limitation of the symptomatic value of
national bank figures.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 583
Moreover, for our purpose we ought to include overdrafts with
deposits, because from the standpoint of the individual firm or household
they are no less available "funds" than deposits. They fluctuated,
however, as deposits,1 only much more strongly, and hence but accentuate
the phenomena under discussion. But they are included in loans and
discounts prior to Dec. 1, 1898. Collateral loans can be separated from
others since 1892. This has not been done, because, as mentioned in
the first chapter, they may and often do serve purposes other than stock
exchange speculation. Concerning time deposits, reasons will be offered
later in this section and in our discussion of postwar developments for
believing that our inability to separate them from demand deposits2 is
less unfortunate than one would think. For the period at present under
discussion time deposits in national (and all commercial) banks were not
very important until 1896. After that date they rose rapidly, and,
although in 1915 they stood at a level which was low when compared with
the figures to which they were to soar in the postwar period, they cannot
be considered as negligible. As a matter of banking practice, however,
it seems plausible that during the prosperity of the late nineties, which
they had to face in fetters that were comparatively tight, national banks
adopted a policy of attracting along with genuine savings also nonsaving
accounts by offering to treat them like saving deposits. If that was so,
we probably lose little by using total deposit figures, without attempting
to correct them for time deposits.
Since, finally, cash in hand and balance with a bank are, normally
and for the individual firm or household, but different forms of the same
thing, the really significant item is cash plus balances rather than balances
alone. Substantially, however, we must, at least for prewar times,
content ourselves with the latter. The only indication we have about
cash in hand is from Money in Circulation, which, besides lacking com-
parability with national bank deposits outside New York City, must be
estimated in extremely hazardous ways. We have a continuous but
highly unreliable series of money in the country outside the Treasury
(specie, bank notes, treasury notes, and so on) from 1813 to date,3
1 Cf. L. W. Hall, Cycles in Banking, p. 74.
2 That inability is not absolute. Professor Mitchell (Business Cycles, 1918, p. 821)
has presented an estimate of time deposits back to 1890. Dr. Thomas's estimate comprises
deposits evidenced by savings passbooks, time certificates of deposits, and postal savings
deposits back to 1834, and thus implies, to be sure, a definition of time deposits that differs
from the usual and official one. But for 1931 these three categories represented about 90
per cent of the total time deposits reported by the Comptroller. And they probably over-
state rather than understate that amount,* which really differed in nature from demand
deposits. For early data relating to Mutual Savings Banks see E. W Keyes, A History of
Mutual Savings Banks in the United States, 1876.
8 Cf. A. Barton Hepburn, History of Currency in the United States, 1915, for early
584 BUSINESS CYCLES
from which the series of money in circulation1 has to be derived by
subtracting money in banks. For the period from 1834 to 1863 the
Comptroller of the Currency recorded both items. After 1900, when
about 90 per cent of the state banks and trust companies, though only
from 20 to 25 per cent of private banks, reported either to state authori-
ties or directly to the Comptroller, we have figures which cannot be far
from the truth and are amenable to fairly convincing corrections. But
before 1834 and during the troubled period from 1864 to 1875 estimation
becomes, to say the least, very risky. For 1875 to 1900 an intermediate
situation prevails. The series of money held by reporting banks is more
significant after than before 1875. But for an estimate of money held
by nonreporting banks we are lacking the indication given since 1900
by the Treasury's estimate of deposits in all banks. There are various
estimates, among which Professor Mitchell's — from the average deposits
of reporting private banks, 1890 to 1911 — seems to hold first place.
We might also extrapolate backward the ratio between deposits in all
and in reporting banks for 1900 to 1914 and apply the result to that
section of the series of money in reporting banks. In any case, however,
it is a matter of estimating from estimates by means of assumptions,
some of which are obviously contrary to fact.
3. If, in order to start from a reasonably simple model, we now focus
attention on items a, b, c, g, and h of our list, we have first to qualify
our expectation as to perfect parallelism between the variations of"
outside clearings and outside individual deposits by taking account of
the phenomena noticed under a and b. In fact, if there are in the system
any deposits at all which do not emerge through loans and, hence, do
not disappear through repayment,2 expectation as to volume of deposits
must, even for the pure model, be supplemented by an expectation as to
data, which can be supplemented by the annual reports published by the Comptroller of
the Currency; see, in particular, those of 1900, 1904, 1920, and 1931. The quality of the
series differs, of course, as between different periods. But it never reaches reliability, the
components of its items, except annual coinage, being rough estimates. For the correction
made in 1907 see Annual Report of the Director of the Mint, 1907.
1 See estimates by Professors Kemmerer, 1897-1904; Irving Fisher, 1896-1909; Wesley
C. Mitchell, 1890-1911; W. I. King, 1880-1920; A. A. Young, 1901-1914. For all these,
as well as for another estimate, 1900-1926, cf. Y. S. Leong, An Estimate of the Amount of
Money Held by the Banks and of the Amount of Money in General Circulation, Jounal
of Political Economy, 1927. The writer acknowledges the help derived from a report by
Dr. C. E. Thomas, who made the attempt to estimate money in circulation for 1813 to 1933.
For our purpose it was not thought necessary to reproduce the chart.
2 Balances that arise from loans need not always be repaid when they are no longer
wanted. And balances which result from customers' selling assets to banks may, although
" owned," disappear in that case, because their owners may prefer to buy assets instead of
keeping them idle. The statement in the text assumes that the possibilities of borrowed
deposits becoming idle and of owned deposits being eliminated are of minor importance
for the argument in hand.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 585
the rate of spending. Obviously, there will be Overspending in pros-
perity, return to normal amounts of (cash plus) balances in recession,
Underspending in depression, and again return to normal habits in
recovery. Clearings should therefore rise more in prosperity and fall
more in depression than balances do. Any increase in recession should
be more marked, the increase in recovery less marked, in the deposit
than in the clearing series. As far as our Chart (XXV) can be trusted at
all, its testimony may on the whole, due attention being paid to inter-
ference of cycles with each other, be invoked in support. It would
be hazardous to go further and to interpret what at first sight gives
the impression of a long-time tendency of the curves to draw apart,
in the light of cyclical effects only. Too many other elements enter into
it. It is, however, permissible to point out that, as far as that impression
is due to the behavior of the curves in the last years of our period, it is
in fact accounted for, in part at least, first by a Juglar and then by the
beginning of a Kondratieff recession.
We might discard the problems incident to the growing divergence of
clearings and deposits by eliminating the "trends" in both series — two
straight-line trends fitted to each of them for the intervals which belong
to the two Kondratieffs would serve the purpose — and then derive a
strongly cyclical series of deviations of clearings divided by deviations of
deposits, which, although open to criticism on the score of the imperfect
comparability of the components, as well as to other objections, could
be said to indicate roughly the variations of the rate of spending in the
shorter cycles. It might serve to interpret certain properties of the
relation between deposits and price level (see the Pulse Charts V, VI,
VII). Recalling the results we finally arrived at about the cyclical
behavior of output, we should, irrespective of overspending and specula-
tive anticipation, expect the price level to rise somewhat less than deposits
in prosperity. Because of overspending it should, irrespective of such
increase of output as actually occurs, rise somewhat more. There is no
reason to believe that these two effects will exactly balance so as to make
price level vary in proportion to the variation of deposits; but they will
balance to some extent. In deep depression again, underspending,
intensifying the effects of decrease of deposits, is partly at least counter-
acted by decrease in output. The rates of spending and of output also
vary in the same sense in revival, and only in recession do they fail to move
together. Since, however, the latter effect is partly suppressed by trend
elimination and, in any case, weak within the shortest cycles, we shall
not share Professor Pigou's astonishment1 at Mr. Carl Snyder's finding
that "deposit velocity" and "trade activity" display a tendency to
1 " Industrial Fluctuations," Chap. XV. On his analysis of English figures see Edie
and Weaver, Velocity of Bank Deposits in England, Journal of Political Economy 1930,
p. 898. Mr. Keynes discussed Mr. Snyder's finding in Treatise on Money, pp. 80-82.
586 BUSINESS CYCLES
neutralize each other's short-time effects on the price level, although we
may be unable to follow Mr. Snyder when he concludes that "neither
variations in velocity nor in trade activity are normally a factor in the
determination of the price level"1: for "trade activity" is obviously not
unconnected with the variations in deposits, which, according to Mr.
Snyder, are — along with the "long-time trend of trade-growth" — the
only factors to act on the price level. The picture presented by English
figures (Chart XXXV, Chap. XIII) is not substantially different.
The material of other countries does not lend itself to a similar experi-
ment, but indications may be had. In France, for instance, the "veloc-
ity" of the net (solde) of the current accounts in the Banque de France,
first studied by P. Des Essars (Journal de la Societe de Statistique de
Paris, April 1895), gives us such an indication, which conforms to
expectation.2
1 Cf. Carl Snyder, The Problem of Monetary and Economic Stability, Quarterly Journal
of Economics, February 1935, p. 189. Chart IV, p. 188, may be referred to as illustrating
the phenomenon under discussion. But it must be born in mind that it cannot do more
than give a very rough idea of it. Outside clearings divided by Mr. Snyder's index of the
general level of prices, corrected for trend and seasonal fluctuations, stand up to 1919 for
output. And total clearings divided by national bank deposits are taken to represent
"velocity" or, as we would say, rate of spending. Neither ratio represents satisfactorily
what it should. The method used to construct the output series, from 1919 on, is not open
to the same objections. But the behavior of the other series, during the years from 1926
to 1930, sufficiently shows that it cannot reflect adequately the rate of spending in the
spheres of consumption and production. Reference should be made to Mr. Snyder's other
papers on the subject, in particular, A New Index of Business Activity, Journal of the
American Statistical Association, March 1924; Deposits Activity as a Measure of Business
Activity, Review of Economic Statistics, October 1924; New Measures of the Relations of
Credit and Trade, Academy of Political Science, Proceedings, January 1930, to which is
appended a bibliographical note. In spite of all that we may have to urge against both
the statistical finding and Mr. Snyder's interpretation of it, the facts remain that the
covariation in the sense outlined in the text between output and "velocity" is definitely
recognizable, and that our analysis goes some way toward supporting the results of his.
The affinity becomes still clearer from his argument on p. 27 of the last-named paper, which
explains the long-run rate of increase in output in terms of what we call growth and evo-
lution without, however, connecting the latter with cyclical fluctuations. It is interesting
to compare Mr. Snyder's investigation with Mr. Holbrook Working's study on the relation
between fluctuations in deposits and in the price level (Bank Deposits as a Forecaster of
the General Level of Wholesale Prices, Review of Economic Statistics, 1925). Mr. Working
finds fairly close covariation (somewhat improved by lagging) between the latter and the
deviations of deposits from a trend of growth, which may be roughly identified with the
trend in output. He does not stress the cyclical variability of "velocity." But since
output as a matter of fact fluctuates cyclically, his result, such as it is, implies that. See,
also, Mr. Snyder's charts on pp. 28 and 30 in New Measures of the Relations of Credit and
Trade.
2 See G. Roulleau, Vitesse de Circulation des Diverses Formes du Stock Mon6taire,
Journal de la Societe de Statistique de Paris, April 1937, chart on p. 9. In order to appraise
the significance of these figures, it is necessary to keep in mind the peculiar position of the
Banque de France in the French economy.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 587
4. Underspending is particularly in evidence in those rare cases
in which deposits decrease absolutely: system expenditure as indicated
by outside clearings then decreases still more (observe the declines to
1885, to 1894, and to 1908; if deposits, see American pulse chart (VI),
increased in 1874 and 1875, this was presumably due to the fact that the
national banking system gained ground and to the immigration of currency
into banks). Such "hoarding" is wholly consequential and, as has been
pointed out above (1, a), fundamentally the same phenomenon as
shrinkage: if business were entirely financed by bank loans and if there
were no other deposits than those created by these loans, there would
be no underspending,1 but all its consequences would still be present
because of the additional shrinkage which then would take the place
of underspending. It does not follow that measures directed against it
are meaningless. Both underspending and nonborrowing are though
consequential yet sufficiently important in themselves and as "causes
of secondary effects" to warrant operating on them directly. Barring
questions of technique, the problem is the same in both cases. Buy-now
campaigns, public works, even stamped money, and so on may help to
break spirals and to relieve depressive situations. Such treatment is,
to be sure, merely dermatological. It also may have, and in general has,
other effects besides the one desired. But it is not futile.
This, however, has little to do with Saving. Inasmuch as the latter
presupposes the existence of owned balances (including cash) which do
not, except in the case of bankruptcy, disappear when not in use, it will
indeed tend to accentuate underspending as against shrinkage of deposits.
Since it is immaterial which happens, there is no point in trying to stimu-
late expenditure by penalizing Saving except this: as far as saving is —
though not necessarily — associated with real investment and real invest-
ment with postponable expenditure on durable producers' goods, savings
are, in fact, more likely to become idle than the sums earmarked for
expenditure on transient consumers' goods of the "necessary" type.
This consideration does not, of course, suffice to "justify" a long-run
policy that aims at transforming the former into the latter, which,
if embarked upon a century ago, would have made it practically impossi-
ble to attain the present standard of life of the masses and which would
logically also have to be applied to sums intended for expenditure on
durable consumers' goods. Moreover, recalling our analysis in Chap.
Ill, Sec. A, and our discussion of individual crises and depressions in
1 To be exact, we ought to say that there will be less of it; for some decrease in activity
will occur also with borrowed deposits. Moreover, there is such a thing as borrowing in
order to be liquid. But its importance can hardly be great, and it is much more likely to
occur in prosperity — when opportunities may be expected to present themselves suddenly —
than in depression, when paying off bank debts or not renewing loans is the most orthodox
thing to do.
588 BUSINESS CYCLES
Chaps. VI and VII, we conclude that depressions may be expected to
be the milder, the more, other things being equal, the expenditure of
preceding prosperities has been financed by saving and accumulation.
But again, as a temporary expedient to be applied when depression has
already set in, such an antisaving attitude, though still more productive
of undesired consequences than simple stimulation of spending, cannot
be called futile.1
Of course, it would change but the form and not the §ubstance of our
argument if we defined saving and accumulation so as to include, or to
coincide with, nonspending. Tautological propositions that would then
follow about the effects of saving in this sense cannot teach anything
about the effects of thrift. That argument does not, however, dispose
of the question what these effects actually are. One point has to be
added, which turns on the relation between saving and credit creation:
although in a system that is beyond the stage of what we have called
the immigration of money into banks, "savings do not create deposits,"
they annihilate deposits whenever they are applied to the repayment of
bank loans. This is in fact one of the major pieces of our mechanism —
repayments out of profits induce autodeflation, or, to put it differently,
accumulating profits ex post finance the real investment, which it was
necessary to effect in order to reap them. If, in addition to this, the
savings of households are borrowed for the same purpose, the process is,
of course, accelerated thereby. Savings thus step in to relieve bank
credit, and in this sense the old theory that it is savings that finance
the expansion of the industrial apparatus comes partly true after all,
even in the presence of credit creation. But they do so with a lag,
which is responsible for a sequence of phenomena that would be absent2
if, instead of stepping in at a later stage, savings financed enterprise
from the start, and that would be less in evidence if they did not step
in at all. We are obviously dealing with a very typical practice.
Nothing is more usual, both in the case of enterprise in our sense and
in the case of induced expansion, than to go ahead on bank credit and
then to "fund" the debt by the issue of stocks and bonds. It is in this
role, rather than in the role of primary source of means, that we would
have introduced saving into our pure model, had we introduced it at
all.
This deflationary effect of — not saving and accumulation as such
but — the application of savings and accumulations to the repayment of
bank loans is what can be adduced in support of oversaving theories and
1 What contradiction may possibly be suspected to exist between the last two sentences
of this paragraph will presently disappear.
2 By this, it will be recalled, we do not mean the cycle itself but certain features which,
though secondary in logic, yet are of prime importance practically, for instance, fluctuations
of the price level which superimpose themselves on and accentuate the fundamental ones.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 589
antisaving policies. Neither the effects of the saving and investment
process within the monetary or the commodity sphere, nor such cyclical
variability as may with justice be attributed to it — the "cyclical release
of saving" — can offer any substantial standing ground for them. But
it may indeed be urged that, as far as savings destroy deposits which would
not become idle, they in fact intensify recession and depression. The
losses incident to these processes will then be greater than they would
be if the same sums had been spent on consumers' goods. And it might
even be held that, thus applied, savings do not serve any social purpose
because, once the task which older theory assigned to them is fulfilled
by credit creation, they seem to come in, as it were, for no other purpose
than to create trouble additional to that which innovation would create
in any case. Here we come nearer, therefore, to the standpoint of
oversaving theories — although it is not any "excess" of saving that is
responsible for this additional trouble — than the general drift of our
argument may have led the reader to expect.
The case which can be made out on these lines is, however, much
weaker1 than it seems. For, although there is no way of determining
exactly the quantitative importance of this particular use of household
savings, the behavior of our series shows that it can hardly ever have been
responsible for absolute fall in deposits or in the sum total of incomes;
and only if it had been, would it have had really serious effects. This is
due to the fact that, as a rule, especially during prosperity but also in
recession, repayment from savers' funds merely serves to set free banks'
facilities for lending, which are then promptly used by other borrowers;
and therefore it seems to contribute but little to the coming about of
the upper turning point, except in price level and interest rates. More-
over, a stabilizing effect of the practice must not be forgotten. On
the one hand, firms which succeed in substituting bonds or shares to
their debt at their banks thereby consolidate their position and thus
restrict the extent of the danger zone. On the other hand, as far as this
is done during prosperity, it tends to bridle excesses and thus to mitigate,
rather than accentuate, reactions. It is different, of course, in recession
and as regards the danger that recession slide off into a spiral, when well-
timed and well-dosed government expenditure could no doubt do much
to counteract what then easily turns into functionless catastrophe.
A complete theory of saving and accumulation could, so the writer
believes, be constructed from the disjecta membra that the reader finds
in almost every chapter of this book. But he frankly admits that its
1 Since analysis is our only aim, there is no need to go into the question as to how far
antisaving recommendations, given certain aims, say, of stabilization, may be said to follow
from the above. Taking the practical question in all its bearing, the writer personally
thinks that there is no case at all for them. But that is for the reader to decide according
to his valuations and according to his confidence in agencies of control.
590 BUSINESS CYCLES
factual complement is far from satisfactory. This is primarily due to a
lack of data, which, for prewar times at least, confronts any student of
the subject, however he may arrange his terminology. As far as it is
also due to the particular definition of saving and accumulation adopted
for the purposes of this book, the obvious reply to possible objections
is that there was no choice. Nothing would have been gained by the
use of a concept that, while lending itself a little better to statistical
evaluation, would be irrelevant to the analysis of the real processes
at work : the insoluble problem of how to extricate the relevant elements
would arise nevertheless. We are, for instance, somewhat better off
with respect to real investment ("capital formation") and to investment
in general, although even these figures cannot be trusted beyond what we
know independently of them.1 But we cannot improve our situation by
1 This must be obvious to anyone who undertakes a critical study of what is still the
standard performance for England and the greater part of the nineteenth century (especially
1865—1885; there is, however, a retrospect that notices the more important attempts back
to Petty, Davenant, and King): R. Giffen's Growth of Capital, 1889. The estimates,
particularly of what he calls accumulation and, within accumulation, "free saving" (i.e.,
the' saving that is invested through the stock exchange), are, even though reasonable,
much too rough to be of use. Later on, we have interesting figures for particular years,
e.g., from the Census of Production, 1907, and also series for individual items. They did
not seem to warrant an attempt at constructing a general series, which the writer hopes,
however, will be materially facilitated by a forthcoming work by Mr. Cairncross.
No estimate of yearly "capital formation" seems possible for Germany, but interesting
samples may be culled from the behavior of companies, one of which will be mentioned
later. To a certain extent, this can be done also for this country, but there is besides a
source that is peculiar to it and has recently been exploited by Mr. Snyder. (See Capital
Supply and National Well-being, American Economic Review, June 1936; also Die Bedeu-
tung des Kapitalangebots ftir den Industrialisierungsprozess, Weltwirtschaftlich.es Archiv,
September 1935.) It is the Census Bureau's data beginning with 1826 on capital invested
in manufacturing — value of plant, cash, inventories, receivables, and so on. The writer
has not been able to satisfy himself whether value of plant means original cost, replacement
cost, or book value, and is also on other grounds disposed to believe that the bureau dis-
continued publication of this information not entirely without reason (see the remarks in
the Xllth Census of Manufactures, Part I, p. xcvi). The relation of the "value of real
capital" to domestic savings must, moreover, be expected to be a particularly distant one in
the case of the United States. Yet he gratefully recognizes the merit of Mr. Snyder's
bold step. The relation of the curve to that of National Income (see op. cit.t chart on
p. 203) is certainly what a theorist would expect. The figures for 1889, 1899, and 1904,
excluding working capital, have been ingeniously used by Professors Douglas and Cobb
(American Economic Review, March 1928, Supplement, p. 139) as stepping stones for an
interpolation — although they are not comparable, both because they do not each time
include the same things and because the mergers greatly influenced valuations — by means
of an index of the quantities and prices of the commodities chiefly entering into capital
goods so as to yield yearly figures. Critical comments crowd upon us (for some of them,
see Professor Slichter's remarks in the discussion of the paper), but results may yet be
nearer to the true contours than one may feel justified in hoping. We shall meet the
subject again, but until we reach the postwar period shall continue to use only those indi-
cators which have already been used in the preceding sections.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 591
defining saving so as to make it equal to or identical with investing; for
this would simply mean begging the most interesting question that arises
in this connection, because it is precisely the relation between those two
distinct processes on which more light would be desirable. A similar
objection estops us from accepting the amount of new security issues
taken by the public as an indicator of saving, which, moreover, is vitiated
by foreign buying and the fact that to a great and varying extent sub-
scriptions are financed by bank loans.
Figures for corporate accumulation are for prewar times available
only in samples, the value of which is reduced by the dependence of sur-
plus on a largely arbitrary variable — depreciation — and by revaluation
of assets.1 Estimates of accumulations by farmers, craftsmen, and
private firms in general, which are ploughed back into their own business,
must be highly unreliable. The greatest disappointment, however,
awaits those who still cling to the idea that at least deposits not subject
to check are, ipso facto, savings. This is most nearly true of the deposits
in mutual and stock savings banks in this country and of the Sparkassen
in Germany. But even neglecting the fact that the Continental saving
institution also serves as the bank of the small man and that, as far as it
does, increase in its deposits indicates not saving but simply transition
to another method of keeping cash, we must not forget that a great
part of those funds is assembled for definite acts of expenditure — a
1 More ambitious attempts would be possible, at least for Germany, if one could
neglect those difficulties and a few others. A very interesting study made by the Institut
flir Konjunkturforschung (Anlagetaetigkeit und langfristige Finanzierung in der Industrie
vor dem Kriege, Vierteljahrshefte zur Konjunkturforschung, 1929), which undertakes to
overcome them, must, however, be mentioned. It deals with 20 industrial companies
only (capital in 1918, 404 million marks), some of them quite young, for 1896 to 1913.
Accumulation is put equal to the sums "reserved" (out of net profits), plus the sums
actually written off, minus "technologically necessary" depreciation. This "necessary"
depreciation is, of course, not above objection, nor is the method by which it was arrived at,
if we may judge from the short sentence describing it; but results are hardly invalidated
thereby. We learn that accumulation (innere Kapitalbildung) varied much as dividends
did, though much more strongly; that in the average of the sample it varied cyclically
between 0.6 and 4.8 per cent of the paid-up capital; that in the average of years (between
1900 and 1913) it was 2.5 per cent of the paid-up capital, which equates it to about 20 per
cent of the total net surplus over cost, excluding necessary depreciation, tantiemes, social
outlay (expenditure from net profits for the benefit of workmen), and sums carried forward
(i.e., of the total of disbursements to shareholders plus accumulation); that these com-
panies on the average raised fresh long-term funds — by the issue of shares and bonds — to
the yearly amount of 7.7 per cent of their paid-up capital (as it stood at the beginning of
each year); and that more than all of these means was regularly invested in plant and
equipment before the end of each prosperity : investing outran long-time financing, occasion-
ally even in depressed years. It must be remembered that the concerns investigated were
highly progressive and, as said before, in part relatively new ones. The sample is, hence,
not representative. But it suffices to correct many a mistaken idea about the invest-
ment process.
BUSINESS CYCLES
holiday, marriage, a cottage, and so on, or less definitely for, say, illness
or a rainy day. Again, the reader may think as he pleases about the
appropriateness of the writer's terminology, which excludes from savings
these sums intended to be spent. But if he includes them, he will have
nevertheless to recognize the fact that the sum total of these "savings"
is currently drawn upon for purposes of consumption by the "savers"
themselves, which makes all the difference.1 This applies still more to
the time deposits in commercial banks, only a fraction of which can
possibly be looked upon as genuine savings. The rest represents under-
spending, temporary investment, and, again, preparation for "bulky"
acts of expenditure, while, as has been pointed out before, a part of the
total does not differ in nature from demand deposits. Finally, the
method of deriving households' savings by deducting expenditure on
consumers* goods from income2 minus taxes and charities, is not available
for the prewar, scarcely even for the postwar, epoch. But even if it
were, we should have to remember that the presence of acts of expenditure
which recur in periods longer than the income period must seriously
impair the significance of the results. In an investigation into the
effects of saving on the economic process and, incidentally, into the
validity of oversaving theories of the business cycle, it would obviously
be absurd to include in savings elements of income earmarked for expendi-
ture on durable goods, such as houses, furniture, and motorcars. Yet
this absurdity is not only difficult to avoid, but its influence must be
expected to increase with time, since in a progressive community those
elements of income will, in general, gain in relative as well as absolute
importance.
Certain facts are nevertheless quite clear. The most important of
them is that the amount of saving in our sense — and, be it repeated,
producing a larger amount by means of a wider definition would not alter
the situations, since the additional elements then included would not
have the same effects — must be much smaller than is commonly supposed.
For with a possible exception on the score of hidden reserves of firms,
all our criticisms of the data tend to reduce it. It is also clear — in fact,
a matter of common experience — that the behavior of firms' accumulation
1 Similar considerations also apply to part of the assets of building and loan associations
and of life insurance companies, which reflect methods of providing for future expenditure
that are alternative to saving.
2 It will be remembered that income in our sense excludes capital gains. Since they
emerge outside of the flow of system expenditure, there is for purposes of business-cycle
study no justification for including them. This has, however, been done in the recent
investigation of the Brookings Institution, and it is to this item that both the fantastically
high estimate arrived at and the rate of increase of "saving" in the later twenties of this
century are due. Spending on consumers' goods of such gains is, of course, dis-saving and
is on a par with financing consumers' expenditure by borrowing. See Chap. XIV.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 593
must be strongly cyclical. Dividends do not in general fluctuate as
much as does net revenue. But since we exclude capital gains, the
behavior of households' saving is more doubtful, because the opportunity
to save afforded by an increase in real income may be counteracted by
optimistic anticipations, as may the effect of a decrease by pessimistic
anticipations. It also varies greatly with groups, countries, times
(see, above, Sec. B). But such indications as we have strongly point
toward steadiness of rate. This is the result which has been arrived at
for postwar Germany by Professor Wagemann,1 and which suggests
itself for this country also, if we may put our trust in the evolution from
1880 of the assets of life insurance companies2 and in the deposits in
mutual and stock savings banks. The latter display a (logarithmically)
straight -line trend from 1840 to 1876 — which, no doubt, is largely a
"special" one in our sense of this term — the small deviations from which
are, however, clearly positive during the prosperities of the fifties and
early seventies. Then we observe decline to 1879 and, from that year
on, again straight-line increase (though of a smaller gradient), inter-
rupted by kinks in 1893, 1908, and 1914. So depression, as well as
prosperity, does assert itself in this series, but, the material being what
it is, the outstanding fact — the steadiness — is the only safe one to
stress.
Again, as in the case of the application of savings to the repayment of
bank loans, it should be noticed that, even if saving and accumulation
involved the "locking up of money/' the consequences usually attributed
to them by antisaving theories would not necessarily follow. In that
case, accumulation from profit would tend to dampen the excesses of
booms exactly as restrictive open-market operations undertaken by
central banks. And the dis-saving during deep depressions — the
depletion of the accumulated surpluses of corporations by deficits and
dividend payments and of the saving deposits of the small savers by
current expenditure — would tend to alleviate them. If we observe so
little of these equilibrating effects, this is because there is so little "locking
up." Also, the importance of the cyclical "release" of savings should
not be overrated. Prosperities are, to be sure, periods of supernormal
real investment; and depressions, of subnormal real investment. Savings
like other funds are active in the former and inactive in the latter. But,
as we have seen, it does not follow that investment must decrease in
recession, for all the demand for funds incident to the conquest of new
economic space then asserts itself, and there is, moreover, that type of
1 See, for example, Einfuehrung in die Konjunkturlehre, 1929, p. 117.
2 See Carl Snyder, op. cit., chart on p. 203. That evolution is almost perfectly fitted
by a (logarithmic) straight line and displays but insignificant fluctuations. Observe the
relation almost amounting to parallelism with Capital Invested in Manufactures.
594
BUSINESS CYCLES
demand which waits upon a fall in interest — demand for dwelling-house
building, in particular.
5. We return again to Chart XXV, in order to study more closely the
relation between customers' balances and loans (and discounts) of
national banks. As has been pointed out in 2, the picture "is of two
series which move closely together, and yet with notable differences"
(A. A. Young, op. cit.9 p. 5). The covariation stands out particularly
well if we eliminate the descriptive trend, an operation that is less
objectionable than it otherwise would be, if we keep within a Kondratieff
.OANS AND DISCOUNTS
TOTAL DEPOSITS
NOTE CIRCULATION
1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911
CHART XXXII. — Germany (see Appendix, p. 1063).
1912 1913
phase, as Professor Young substantially does in his chart 15 F (ibid.,
p. 27), l which covers the years from 1901 to 1914. But the essential
fact so important for the understanding of prewar banking at least —
viz.j that the movement of loans dominates the movement of deposits —
is obvious in any case. Nor is this covariation confined to those shorter
fluctuations which are traditionally recognized as "cycles." Of course
the longer the period we survey, the more extraneous elements will
assert themselves, and we shall not be surprised to find that covariation
is particularly close in the course of Kitchins and Juglars. But the
phases of the two Kondratieffs are also recognizable and influence both
series similarly. Chart XXXII presents the German case.2
1 Comparison is there with Net Deposits, but this does not matter for cyclical, although
it would for seasonal, movements; cf., same chart B.
* In perusing that chart, it is necessary to keep in mind that total deposits do not mean
exactly the same thing or hold, as far as they do mean a comparable magnitude, the same
relative position as in the United States. It must also be remembered that deposit banking
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 595
The covariation between loans and deposits must of course be dis-
turbed by member banks* investments. In national banks outside
New York City they fell from a level of roughly 10 per cent of loans and
discounts at the beginning of the series to a minimum of 24 millions
(Feb. 28, 1873), but within our period they reached 898 millions, or
about one-sixth of loans and discounts (June 30, 1914), the increase from
1897 to 1902 helping to push the deposit up to the loan series. This
subject belongs to the sphere of banking practice and will be taken up
in Chap. XIII; but we may note at once that the impression of an erratic
behavior of investments in the course of cycles is, apart from the special
trend that is clearly present in them,1 due to the conflict of two tendencies :
the one which clearly prevails, at least in the shorter fluctuations (see
A. A. Young's chart 15 H, op. cit., p. 27), and works toward negative
association of investments with loans, the banks employing in slack
times idle funds in the purchase of assets, thereby creating idle deposits
that help to steady the deposit series ; and another and weaker one, which
proceeds from firms and individuals' liquidating their holdings of bonds —
when in busy times they have more profitable uses for their funds —
and which works towards positive association with loans and intensifies
fluctuations of deposits.
But there are other factors which will enforce deviation from the
fundamental parallelism of loans and balances. One is variation in the
capital and surplus item. In national banks outside New York City
this item grew from less than 400 millions in 1867 to over 1^ billions
in 1914. It clearly displays the influence of the cyclical process, following
as it does on a trend of smaller gradient2 the movement of earning assets
(see A. A. Young, op. cit., Chart 2 on p. 5 and Chart 9 on p. 22). Another
factor is variation in original deposits. The spread of banking habits
and the consequent immigration into banks of money that used to
circulate outside of them would in itself suffice to impart another special
trend to the loan-deposit ratio. Apart from this, it is easy to realize
that original deposits will also fluctuate cyclically and that they will
reflect the influence of gold movements, although in the latter respect
it is interesting to note that neither the deposit nor the loan series behaves
very differently before and after the intrusion of the new gold. If we
had no other information than that conveyed by those two curves,
in general and the business of the banks that contributed the figures in particular, rapidly
gained ground during that period. This partly accounts for the steep gradient from 1903
to 1906 and again in 1910. In order to bring this out, the series showing the development
of note circulation has been added.
1 See infra, Chart XXXIV.
2 The fact that that gradient is smaller than the gradient of the earning asset curve
reflects, in part, the relative decrease of hand-to-hand circulation (see below).
596 BUSINESS CYCLES
we should hardly be able to infer that something had happened to alter
the monetary data of our process. Finally, as long as coins and notes
circulate at all, and especially if farmers, craftsmen, wage earners, and
their retailers live to a significant extent outside the banking sphere, the
total of balances will, other things being equal, tend to increase and
decrease cyclically by less than loans,1 because cash is gradually with-
drawn from every original or newly created deposit when business is
brisk and, however quickly redeposited, dwells in greater amounts
outside the banks in prosperity than in depression.
These considerations suffice to explain why, on the one hand, the loan-
deposit ratio is itself a cyclical variable, and why, on the other hand, it
is not always easy to interpret its fluctuations or its trend. Mainly
because of the predominantly inverse association between investment and
loans and of the cyclical drain and reflux of cash, that ratio moves with
an understandable lag in the same direction as its constituents.2 But
several circumstances besides those mentioned above — depletion of
deposits by hoarding from distrust in banks, for instance, or legislative
changes — combine to interfere with that rule. We may now inspect
Chart XXXIII, on which circulation (notes outstanding) has been added
to individual deposits, both on the theory that bank notes are funda-
mentally the same kind of thing as customers' balances and on the much
more doubtful theory that the variation in their amount is indicative
also of short variations in the amount of legal tender in the hands of
the public, therefore, of the total of money outside the Treasury and
all banks. Over time, of course, the circulation of national banks
is too much under the influence of factors peculiar to them to be of
much significance for us.
As far as the variability of the loan-deposit ratio is due to the cyclical
depletion and repletion of deposits by cash moving out and in, the ratio
between money in circulation and deposits must also be a cyclical (and
seasonal) variable. So it is, of course, and the explicit or implicit
assumption characteristic of several well-known versions of the quantity
theory, viz., that it is constant, can only hold for a perfectly stationary
state. In a nonstationary system this assumption cannot even hold
1 Actually, deposits fluctuated more than loans, though not so much more than in New
York banks. This is not difficult to understand. It must be remembered, however, that
what we have before us are the figures not of a complete system but only of a segment.
8 According to Professor W. M. Persons, Cyclical Fluctuations of the Ratio of Bank
Loans to Deposits, 1867-1914, Review of Economic Statistics, October 1924, this lag is
about six months. While deposits and loans themselves fluctuate concurrently with each
other and the Harvard Business Curve, their ratio thus fluctuates concurrently with the
Harvard Money Curve. This is as we should expect. Professor Persons, however, uses
the ratio we are discussing only for the last prewar decade. We shall not be astonished to
learn that those relations are less clearly discernible before 1900.
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 597
over time — say, for successive neighborhoods of equilibrium — for our
process changes the relative importance of the transactions directly
settled by cash : it is the most powerful factor responsible for the immigra-
tion of money into banks, i.e., the spread of the habit of paying by check,
which in turn is the most powerful factor in the expansion of the lending
1870 1880 1890 1900 19)0 1913
CHAHT XXXIII.— United States (see Appendix, p. 1063). -
facilities of banks. This, first, helps to account for the fact that the
loan-deposit ratio has been constantly falling over time.1 While this
ratio can be observed in the (outside) national bank data, measurement
of the second ratio — that between currency in circulation outside the
1 It averaged about 160 per cent in the seventies and fell in every successive decade,
until it practically reached 100 per cent by 1910. This is, of course, not only due to
deposits losing less and less to outside circulation, nor can the sharp jerk from 1879 to 1881
be explained in this way. A few figures should be added: Between 1875 and 1914 loans and
discount rose about seven fold, individual deposits more than elevenfold (11.8). "Lawful
Money Held" by the same banks was 76 millions on Mar. 1, 1875, and 642 millions on
Mar. 4, 1914; Capital and Surplus on the same dates were respectively 536 and 1,539.
598 BUSINESS CYCLES
Treasury and all banks and total deposits — is, as we have seen, a more
doubtful matter. But at least from 1890 on, these data cannot be so
faulty as to shake our confidence in the strongly falling trend they
display.1 Third, the ratio of outside circulation to loans plus investments
also has an obvious significance of its own. Mr. Carl Snyder tried to
estimate it2 from an estimated series of outside money and from national
bank loans and discounts "adjusted upward to the level of all commercial
banks in 1913." But whatever might be urged against this, the salient
fact is again beyond doubt. Since the fall in all three ratios is sub-
stantially the result of the cyclical process of evolution, we may speak of
result trends.
6. The reader should once more refer to Chart XXV and then inspect
Chart XXXIV, which, although very inadequately, presents the behavior
of indicators of what is, in one sense or another, called Investment. The
graph of member banks' investments as represented by the investments
of national banks outside New York City has been inserted merely for the
purpose of showing — what would be more obvious if a descriptive trend
had been eliminated — that their fluctuations are about as much inversely
related to production of industrial equipment as they are to loans and
discounts, that is to say, predominantly yet not entirely so. There is
again more than a suggestion, as there was in the comparison of loans
with deposits, to the effect that selling assets to banks is to a significant
extent resorted to by individuals and firms as a method of providing
funds for industrial expansion. Building Permits do not merely reflect
investment — since the building of residences for one's own use is not
investment in our sense — and behave as erratically as such a composite
of consumers' and investment goods should.3 All affinity with the move-
1 C/. Professor Angell's ratio h on his Chart II in Behavior of Money, 1936. The deposit
figures used include all deposits in reporting banks, also in postal savings banks, and both
government and interbank deposits. The float is, however, deducted. But the present
writer is not able to agree with his statements that outside currency, circulating deposits
(deposits subject to check — bankers' balances — clearing house exchanges) and the ratios
of outside currency and circulating, as well as total, deposits do not display clear cycles or
that there is no evidence of direct causal connection between currency and deposits. On
the contrary, all our fluctuations seem to show very well: the downward sweep of ratio h
from 24 per cent in 1890 to 10 per cent in 1914 is broken in every prosperity, intensified in
every recession or depression. It should be added that that downward tendency is
exaggerated because of the decreasing importance of the money held by nonreporting
banks.
2 "The Problem of Monetary and Economic Stability," Quarterly Journal of Economics,
February 1935; see chart on p. 192.
•The Berlin Institute's study on industrial investment and financing quoted in a
previous note presents a chart which compares the increase in industrial buildings (Fabriken
und Werkstatten) in 32 cities from 1895 to 1918 with pig-iron consumption, trend being
eliminated. The covariation is striking, with industrial building construction slightly
EXPENDITURE, WAGES, CUSTOMERS' BALANCES 599
meats of loans and discounts is not wanting, however, especially until
the end of the second Kondratieff, although it reveals itself more readily
to an analysis of underlying business situations than it would to formal
measurement. The covariation of New Security Listings — intended to
indicate roughly fluctuation of households' investment — with loans and
discounts is, except in the early nineties, fully as strong as we can expect
LOANS AND DISCOUNTS
NATIONAL BANKS
OUTSIDE NEW YORK
INVESTMENTS /
NATIONAL /
BANKS OUTSIDE
NEW YORK /
1870 75 I860 85 1890 95 1900 05 1910 13
CHART XXXIV.— United States (see Appendix, p. 1063).
if we bear in mind the difference in the "temperaments" of the two series
and the difference in structure of the "resonators" of which they are the
products. Firms' real investment is again represented by an index of
the production of industrial equipment. Its graph consistently draws
away from that of loans and discounts, but otherwise varies with the
latter, though again in the manner agreeable to its temperament — a
strong jerk in the one often corresponding to a mere kink in the other.
Both series display the two Kondratieffs in much the same way, clearly
but almost consistently preceding. Industrial building would on this showing turn out
to be an even better index — or forecaster — than pig-iron consumption,
600 BUSINESS CYCLES
show the Juglars, move together in the Kitchins. The reader should
satisfy himself that such exceptions as seem to occur are no more than
understandable lags. Even the fact known to ail of us from everyday
experience, that there is such a thing as emergency borrowing and that
every well-connected concern can go pretty far in this direction, need
hardly ever be invoked.
The time shape of loans, therefore, corresponds — in the long and
in the short run ; in trend and in all the cycles — as perfectly to expectation
from our model as the time shape of deposits that it controls. But
again, if we go on to say that their behavior is illustrated and explained
by the behavior of the pig-iron consumption or equipment-production
series, we are adding something to the testimony of our charts. Repeat-
ing the corresponding argument submitted in the case of system expendi-
ture, we have once more to recognize that, although perfectly safe
and supported by common opinion, the proposition that investment is in
prosperity, though not in revival, the propeller of general business activity
and hence controls the behavior of loans and balances, is such an addition
from other evidence, an addition moreover which is not so sheltered
by triviality as the proposition that loans control deposits and not
deposits, loans.1 Less safe, however, is the further addition that loans
substantially contribute to the financing of real investment, and still
less the final one that they are the — logically primary — source of real
investment incident to innovation. As far as time-series evidence goes,
it would be just as possible to say that it is the movement of loans and
deposits, itself determined by monetary conditions and policies, which
induces and controls real investment, or that initiative is not with entre-
preneurs but with banks.
This issue can be decided only by theoretical and historical analysis
of the capitalist process, such as has been attempted in earlier chapters,
where it has also been pointed out that the fact that loans and discounts
mainly finance current operations — in each cycle perhaps preponderately
those of the Secondary Wave — and even consumers' expenditure, con-
tradicts our interpretation in appearance only, and that the extent to
which they actually serve the purpose of financing long-time commit-
ments and in particular industrial plant is likely to be underrated,
because that purpose hides itself. We have seen by such outstanding
examples as the rise of the automobile industry that it need not be the
entrepreneur who borrows from a bank, but that the burden can be
shifted to the firms who furnish the materials or market the product;
or that the stock exchange speculator or the investor who takes new issues
1 Of course, in another sense that proposition is not trivial, viz., in the sense that it is
loans and not member banks' investment which dominated the behavior of deposits in the
period under discussion,
EXPENDITUKE, WAGES, CUSTOMERS* BALANCES 601
may, and very often does, borrow instead of the entrepreneur. But
we have also seen in the historical chapters plenty of indications of bank
loans' being directly used for the purposes of innovation, especially in
this country,1 less so in Germany, least of all in England. In fact, if the
reader will refer to the list of the methods of financing system expendi-
ture at the beginning of this section, he will have no difficulty in satis-
fying himself that it must be so. Though borrowing from banks is not
the only way of — directly or indirectly — financing innovation, and
though bank loans also serve many other purposes, our schema cannot be
so very far from practice after all; for most of those methods are either
inadequate or not available for the purpose of innovation which starts
from scratch.
Glancing over our list of methods of financing, we readily see that
deflection, overspending, and the use of temporarily available funds
will neither be open to a new firm nor, in general, prove sufficient to
finance large-scale plant and equipment. Accumulation and saving
provide the means for consolidating rather than for building up industrial
ventures. In any case, they owe their actual importance only to previous
entrepreneurial success and serve, besides repayment of debt, induced
expansion rather than the creation of entirely new things, although the
latter function has always played a considerable role from the times of
primitive, through the times of competitive, to the times of "trustified'*
capitalism. The same applies to selling assets to non-banks. Neglecting
items of secondary importance, we are thus in fact left with selling
assets to, and borrowing from, banks, as influenced by borrowing from,
and lending to, foreign countries. A rough correspondence between
innovation and loans, still more between innovation and deposits, is
hence not as improbable as it might seem at first sight.
1 Some further indications may be noticed, however. Mr. Moulton, bolder than the
present writer, has actually presented an estimate of the relative importance of the advances
made by banks in 1916, that went to "fixed capital operations," and found them to be 55 per
cent in the case of national, 62.1 per cent in the case of state banks, and 68.2 per cent in
the case of trust companies (Journal of Political Economy, June 1918). The responsibility
is Mr. Moulton's. Or see C. O. Hardy and Jacob Viner, Report (to the Treasury Depart-
ments), on the Availability of Bank Credit in the Seventh Federal Reserve District, which
states that many short-term loans (not to speak of mortgage loans, of which those on non-
farm land are certainly relevant to our argument), being in fact renewed indefinitely, are
used by borrowers for the purpose of financing expenditure on plant and equipment. Part
of these loans must finance enterprise in our sense.
CHAPTER XII
The Rate of Interest1
A. Earlier Argument Resumed. — Resuming our argument about the
rate of interest where we left off in Chap. Ill, Sec. E, and recalling that
we look upon it, on the one hand, as a premium on present over future
means of payment or "balances" and, on the other hand, as a coefficient
of tension in the system, we will first of all develop some of the aspects
of the theory there presented. It is convenient to take the first step
under cover of all the simplifications of the pure model: once more nobody
is to borrow except entrepreneurs, nobody is to lend except member banks
which create the required "funds" ad hoc, all other business — consisting
only of purchases of producers' goods and sales of consumers' goods —
being financed from previous receipts. Excluding, moreover, for the
moment what, in Chap. Ill, Sec. E, has been defined as the Open and the
Central Markets, we have at least the comfort of facing one rate of
interest only, which no doubt may differ as wages do in different parts of
the system, but can, nevertheless, be used in the same way as the wage
rate. While all these simplifications will be dropped within this chapter,
we defer to the next the consideration of international capital movements
and as much as possible also of everything that happens within the walls
of member banks' committee rooms beyond acceptance or refusal of
applications submitted by entrepreneurs on their own initiative. In
other words, in this chapter we deal primarily with closed domains and
with a "passive" banking system, in order to see how great an expanse
of observed fact can be covered within these restrictions.
1. In the absence of consumers' time preference, innovation, or the
expectation of profits in our sense of the term, is now the only factor to
produce, in an otherwise stationary and undisturbed world, a positive pre-
mium on present balances or to make present dollars worth more than the
same amount of future dollars. Looking at any amount of dollars plus
the profit an entrepreneur hopes to realize by means of it as bis "demand
price" for that amount, we seem to be able to build a curve or schedule
of entrepreneurial demand for present dollars, the ordinates of which are
1 Reference should be made again to Theory of Economic Development, Chap. V. But
the reader is once more reminded that many of our propositions are not dependent upon
the writer's theory of interest,
602
THE RATE OF INTEREST 603
expressed in terms of future dollars, and to conclude in the familiar way
that interest will equal marginal profit, i.e., not merely the profit of the
marginal entrepreneur but, since we may postulate that every entre-
preneur can continuously vary the size of his commitment, the marginal
increment of prospective profit for all. But such a construction means
very little in our case, although it has the merit of showing the nature of
the link between our monetary theory of interest and the general theory
of the industrial process. We have not only to recognize, as we always
have, that no causal significance attaches to marginal values but also
that all the significance they have lies in the setting up of equilibrium
conditions. Our case, however, is essentially one of disequilibrium and
the circumstances that determine marginal profits are being changed,
and known to be changing, every moment by the very process we are
trying to describe.1 What we can hope to get at, therefore, is at best a
rough equivalence between interest and marginal prospects of profits
or a condition not of equilibrium in the full and proper sense of the term
but of adapted variation within our process, i.e., of covariation such as will
not by itself enforce additional adaptation of other elements of the system.
This is all we have to offer in place of the various definitions of the con-
cept, Equilibrium Rate of Interest.2 The true equilibrium rate, i.e.,
the rate that would obtain in a stationary process perfectly equilibrated
in all its elements and displaying no tension at all, would as we know be
zero.3
Even if we decide to neglect these difficulties and to speak of such a
demand schedule after all, we have still to recognize that, more than any
other, it shifts violently in the cycle. This trivial consequence of any
1 That the analogy with the demand schedule for a factor is treacherous is also obvious
from the fact that well-behaved demand schedules presuppose invariant production func-
tions. In our case, production functions are being altered or new ones are being built up.
Moreover, capital in our sense, not being a factor of production, enters neither the existing
ones nor the new production functions.
2 As pointed out in Chap. Ill, Sec. E, that condition to some extent stands in the place
of the two conditions that are required in the Wicksellian system, viz., first, that the natural
rate of interest should be in equilibrium — for example, that the rate of time preference
should equal the rate of technological superiority of present over future units of original
means of production — and, second, that the monetary rate should equal that natural rate.
For a development (and improvement) of this theory see G. Myrdal, Der Gleichgewichts-
begriff als Instrument der geldtheoretischen Analyse (Beitrage zur Geldtheorie, ed. by
F. A. v. Hayek, 1933). From theories of the equilibrium rate of interest we should distin-
guish theories of the equilibrating rate of interest, for it is obvious that the two do not
necessarily coincide. A rate that fulfills the ordinary conditions of equilibrium, for
instance, a rate that clears the market of savings, might, nevertheless, be the cause of
disturbances, for instance, through its action on price level. Whoever holds that it would
and that another rate would not had better state separately the conditions that define the
two.
8 But, once more, the author does not here insist on this proposition.
604 BUSINESS CYCLES
analysis that admits any relation at all between interest and profits in
our sense is also so obvious to any observer of business life that it should
be superfluous to state it. Yet it is constantly overlooked. The
exaggerated ideas that many economists entertain — though perhaps
somewhat less so now than they did 10 years ago — about the effectiveness
of a reduction of the rate of interest in inducing investment is under-
standable only on the hypothesis that they assume — besides a full measure
of ceteris paribus — that the demand schedule for funds is substantially
invariant, at least for periods as long as Kitchins and perhaps as long as
Juglars. The familiar calculations first stressed by Wicksell of the
tremendous difference it makes to railroad building whether interest is
1 per cent higher or lower, on the one hand, and the frequent recurrence
of complaints about the failure of interest to fall adequately in recession
and depression, on the other, which are so curiously at variance with the
fact that new investment and even replacement is primarily — we have
met exceptions — associated with relatively high and rising money rates,
prove better than anything else how completely most economists still
think in terms of expansion within unchanging production functions and
how reluctant they are to recognize that in the cyclical process shifts and
distortions of the schedule of demand for balances are much more impor-
tant than movements along it. There are situations in which zero
interest would entirely fail to call forth any additional demand.
2. These shifts of the demand schedule for balances are, of course,
intensified by the operations that constitute the Secondary Wave. This
phenomenon we have already met with in the case of wages, but it asserts
itself much more directly and much more strongly here. There is, how-
ever, another kind of shift which still more seriously interferes with the
application of the demand-supply apparatus whenever there is full
employment of resources. Then a net increase in balances created in
response to business demand1 enforces a rise in price level, induces
expectations of further rise, and for both reasons further increase in
demand for balances, so that that demand schedule tends to shift also in
function of the amount of, and increase in, the balances themselves,
and in consequence completely ceases to be independent of "quantity of
commodity," as a demand schedule should be. If entrepreneurs really
were the only borrowers, we should not commit any great error by
neglecting this effect. For, as we have seen, entrepreneurial action is not
typically taken on an expectation of a rise in price level or under the
influence of a price level that has risen already. Moreover, although
entrepreneurs may not be aware of the falling result trend in price level,
they cannot but be aware of the fall to be expected in the prices of their
1 It will be observed that this is not equivalent to saying that any net increase in
balances created will, even with full employment of resources, necessarily have that effect.
THE RATE OF INTEREST 605
own products, which they themselves intend to bring about in order to
use their innovations to the full and/or to undersell old firms; and this
will in general suffice to prevent them from acting on a hypothesis to the
effect that prices will go on rising forever. But actually, since there is at
any time in prosperity a large mass of windfall gains in prospect, the
general willingness of business to pay interest will, both because of the
mechanical effect of rising prices on the money cost of doing business and
because of extrapolating anticipations, increase — and correspondingly
fall in depression — by much more than it otherwise would. The rate of
interest will hence no longer be explainable by the fundamental factor
alone, not only because there are other borrowers, but also because many
of them strongly react to the rate of change of prices.
It follows that what we have termed the Adapted Rate must now be
defined with reference to a "demand schedule" for balances in terms of
expected windfalls, behind which the entrepreneurial profits that "ignite"
the process may at times entirely disappear. This rate will differ, in
prosperity positively, in depression negatively, from the one that would
be appropriate to entrepreneurs' profit expectations over time, and
change literally every hour. But it will, in general, also differ from what
we may call the equilibrating rates, i.e., the rates that would tend to
counteract disturbance. For the adapted rate will not do that as long
as prosperity is in full swing and as long as the balances that are being
created in every point of time incessantly change the conditions under
which the corresponding loans were applied for.1 This fact underlies all
arguments for an active policy of banker's banks in prosperity and for
"punitive" rates. In a competitive banking system member banks will
charge the adapted rate and then "catch up" only when windfalls,
as well as profits, begin to fall in the course of the cyclical process.
This being one of the causes of what is usually referred to as the Lag
in Rate of Interest, we will take the opportunity of disposing of it at once.
Discarding the spurious lag resulting from the failure of students of the
cycle to distinguish between revivals and prosperities, and also the
frictional lag that is naturally incident to a relation such as establishes
itself between a bank and its customers and is intensified on the upgrade
by the pressure of public opinion2 and on the downgrade by risks and
fears, we have, besides the one mentioned in the preceding paragraph,
a fundamental reason for expecting that interest will tend to lag behind
other cyclical symptoms. It can best be described in terms of our pure
model. There prosperity starts from zero loans, so that the lending
1 This may be called a cumulative process and bears some resemblance with what is
known as the Wicksellian cumulative process. Possibly we are visualizing the same facts
as Wicksell. But the two arguments themselves have little in common.
2 Frictional lag is, however, almost entirely absent from open market rates, which in
fact are the most active ones by far.
606 BUSINESS CYCLES
power of banks is entirely unused, while, at least in perfect equilibrium
of perfect competition, all the factors of production are used to optimum
points. In reality this is not so, of course, but our abstraction still
serves to show why full employment does not imply that full use be made
of the existing machine for the manufacture of credit and why this
machine will in general be underutilized in a neighborhood of equilibrium.
Now in a world conforming to the pure model, interest would also start
from zero and hence be one of the first elements of the system to move.
But since there is actually in every neighborhood of equilibrium a positive
rate at which banks will in general be glad to expand their loans, that
rate will not rise immediately.
3. Thus the demand curve for balances is, even with greatly simpli-
fying assumptions, an instrument of doubtful value. But the value of the
corresponding supply curve is more doubtful still. Certain difficulties
would remain, even if we could interpret it as the supply curve of savings,
as older doctrine did. But at least there would then be some kind of cost
curve to derive it from and hence at least some definite meaning to it.
Since, however, this is out of the question, and since it would similarly
not help us much if we tried to build a bankers' cost curve in terms of
increasing risks, the natural thing to do seems to be to drop "supply " and
to fall back upon "quantity existing."1 But this too would be entirely
unrealistic, because, as has been explained in Chap. Ill, Sec. D, there is
never any such thing as a definite quantity of bank accommodation avail-
able, not even in a perfectly competitive banking system — if every bank
moves pari passu with the others — while in an imperfectly competitive
one the individual bank has a large space to maneuver in. Boundaries
are not lacking, but we have seen how very elastic they are. Hence, it is
necessary to recognize an element of indeterminateness in the problem of
interest, which is precisely the theoretical reason why regulating the
money market differs in kind from regulating any commodity market.
We may now insert the two other major items that stand alongside
the created balances — savings,2 and funds that are temporarily available,
either because they are being assembled for lumpy acts of expenditure or
because their owners have decided to restrict operations. We recall our
concept of Temporary Investment and note for future reference that it
plays a role in the theory of the rate in the open market, where those
1 It should be obvious without further explanation that the sum total of deposits would
not measure that quantity or the "supply of bank credit" any more than bank loans
measure the "demand" for it.
2 The rate of savings could be made a function of the rate of interest only by very severe
ceteris paribus assumptions. Even then it would, as everyone now admits, not be a simple
or a monotonically increasing function. But this is no reason to deny the existence of a
functional relation. The same applies to the inverse relation which, in Chap. Ill, Sec. A,
we were able to deal with so simply because we assumed otherwise stationary conditions.
THE RATE OF INTEREST 607
funds join forces with the otherwise unused means of member
banks.
Similarly, we may now insert the other major items of total borrowing.
These consist, first, in the amount of borrowing that proceeds from the
Secondary Wave already mentioned, the requirements of induced and
adaptive expansion, and also those of current business, since as a matter
of fact current business, too, is partly transacted on a credit basis.
And, in the second place, there is the vast amount of consumers' and
quasi-consumers' borrowing, in which we include the emergency borrow-
ing by firms; the cases (which are numerous, especially in the agrarian
sector) of borrowing which, while in name productive, mainly serves pur-
poses of consumption; and finally, the borrowing of public and private
households. The cyclical behavior of some of these items is doubtful.
We have, for instance, seen reasons to believe (Chap. XI, Sec. B) that at
times private households borrow more, both in the form of installment
contracts or open accounts and in the form of loans, in prosperity than
they do in any not too prolonged depression. But this may not be so
true of prewar times and Europe as it is of postwar times and the United
States. Government borrowing, as far as it is incident to war and
preparation for war, is erratic. We might, however, also expect to see a
cyclical component, owing to the fact that revenue falls off in deep
depression, even though depression expenditure did not in our epoch
play the role it came to play after the World War. This expectation is
not entirely disappointed, witness the English case in the last years of the
Melbourne-Russell government. But in the times of "sound" fiscal
policies such cases were the exception, not the rule. Goschen's budgets
more than balanced in the midst of depression. The heavy borrowing
of the German Empire occurred in the prosperity phase of the third
Kondratieff.
It will be observed that some of the remaining items tend to intensify,
others to mitigate, those cyclical variations in interest rate that we would
expect from the entrepreneurial impulse alone. But it will be also
observed that with the exception of the requirements for induced expan-
sion or adaptation, which will tend to keep up the rate of interest in
recession, there is little reason to expect that they will substantially alter
its time shape.
4. From its source interest spreads, as we have seen, all over the sys-
tem. A premium on present dollars in any sector is sufficient to enforce a
general premium in all. Thus interest intrudes into every transaction,
calculation, and valuation, turns time into a cost factor, and becomes
that subtle and omnipresent entity that acts on and reacts to everything
and is so difficult to trace in all its protean forms. Every unit of currency
or deposits, wherever placed or circulating, has, in order to stay where
608 BUSINESS CYCLES
it is or to go on in its circuit, to resist a pull toward the money market,
which at the margin is measured by the rate of interest. This point may
be illustrated by means of a Wicksteedian demand curve if, for the
moment and for the purpose of illustration only, we assume that there is
such a thing as a definite quantity of (cash and) balances in existence,
including the amount, also assumed to be given at any point of time,
which banks can technically create beyond what they actually have
created at that given point of time. The Wicksteedian demand curve
refers price not to the quantity of a commodity that buyers are willing
to take at that price but to existing quantity, i.e., that quantity plus
the quantity which owners keep at that price. Accordingly, we may look
upon every unit of actual or potential balances as " offered" on the money
market and taken from it either by its owner or someone else. The
owner must then be thought of as paying interest to himself, either in
the form of some element of return if he uses his money in his business,
or in the form of some satisfaction (equivalent to the loss of interest
involved) if he does not.1 But apart from being applicable to the case of
perfect competition only, this schema presupposes a string of assumptions
that are entirely inadmissible in the case of money. Although it may be
used to clarify that delicate point at which our theory proceeds from the
analysis of the logical sources of interest to the analysis of interest as an
all-pervading phenomenon, it should not be taken too seriously.
In this sense interest may indeed be said to hold a central position
in the system. But again, precisely because it does, it is easy to slide
into exaggerations of the influence unilaterally exerted by it and to forget
that a central position does not imply a key position, particularly in the
cyclical process of evolution. Just now, however, two cognate errors
call for our attention.2 With the growth of capitalist mentality, an
obviously useful habit has developed, the beginnings of which are in Ger-
many, for instance, observable since the fourteenth century, of expressing
any return, except returns to personal services, as a percentage of a
capital value. But it is an error to conclude that these returns are
turned into interest thereby or, worse still, that they are the basic fact
about interest and, barring temporary discrepancies, fundamentally
determine the money rate. What any durable product yields — no matter
whether a consumers' or a producers' good, though rational calculation
is understandably more in evidence with respect to the latter — once it has
1 Viewed from this standpoint, interest could also be defined as the price that it is
necessary to pay to a holder of actual or potential balances in order to induce him to part
with "his money." Thus we meet, for a moment and under very restrictive assumptions,
the concept of interest which has, in the General Theory of Employment, Interest and
Money, been adopted by Mr. Keynes. But the point of tangency between our argument
and his is not more obvious than the 'divergence of the curves.
2 For a full discussion of what follows see Theory of Economic Development, Chap. V.
THE RATE OF INTEREST
come into existence, is and remains (temporary) quasi-rent,
may express it. It is neither the same thing as interest nor i
equal to interest. The other error consists in the interpretation
discounting process by which the capital value of a durable good is arrived
at and which indeed establishes a relation of mutual dependence — ideally,
equality at the margin — between quasi-rents and interest. But it does
so by applying a logically prior monetary rate of interest — that monetary
rate which is expected to prevail during the life of the good1 in the sector
within the horizon of the firm or household concerned — to the sequence
of the expected installments of quasi-rent evaluated in money, and not by
virtue of an interest logically independent of money. It, hence, pre-
supposes the monetary rate, instead of controlling it.
Of all the ramifications of this principle of evaluation and all the
qualifications that it calls for, only two need to be touched upon. First,
we have again, as in the case of the demand schedule for balances, to
bear in mind that the demand schedule for durable goods — -which, though
doubtful enough, is nevertheless more like what a demand schedule should
be — is subject to very strong shifts and distortions, which, in the course
of cyclical phases, are more important than any movements along it.
Professor Moore's rising-demand curves for steel and other products that
enter into durable goods (Chap. X, Sec. A.) give a good idea of how com-
pletely the latter are overshadowed by the former. This, of course, is
only what we should expect. But it practically implies an analogous
proposition about the influence of variations of interest. Although this
does not impair the logical validity of the theorem that values of such
goods will, ceteris paribus, rise and their production become more profit-
able if a fall in the rate of interest occurs or is anticipated, it does impair
its relevance to our subject. The outstanding fact that is really relevant
to our subject is the positive association of rate of interest and value of
capital goods. Our motive for insisting on it is, again, that many state-
ments in older as well as more recent literature are understandable only on
the hypothesis that it has been overlooked.2
1 That life is not a (technological) datum but a variable of the problem; but this, like
other minor points, need not concern us here.
2 The objection which might be raised against the above argument, viz., that it is not the
money or the natural rate that counts, but the difference between the two, misses the point,
or, if natural rate is but another name for profits in our sense, merely reformulates it.
But the above argument should not be interpreted to mean that financial editors are wrong
in hailing a falling rate of interest as a hopeful symptom in a difficult situation. Although
its causal importance may be next to nothing — we do not evaluate it at zero, however —
it is a symptom of decrease of strain, of progressing adaptation and liquidation, and in
many cases it is, and especially was in the nineteenth century when panics were allowed to
have the effect of forcing up the rate, a sign of the worst being over. If the businessman
clamors for cheap money, that means as much as and not more than his clamoring for cheap
labor.
610 BUSINESS CYCLES
A second point turns on the fact that, even where it is possible to
identify the rate1 at which it is rational for a firm or a household to dis-
count a given series of quasi-rents, we in general find a considerable dis-
discrepancy between the result and actual values of durable goods or
going concerns. There are many obvious reasons, both sociological and
other, which need not detain us, why, e.g., rural real estate at times sells
much above the price our principle would indicate. But one of these
reasons is especially important for the subject in hand* viz., the risks and
chances of depreciation and appreciation incident to ownership of an
individual piece of property. Whoever acquires a durable good commits
himself. He chooses one out of an innumerable host of possible sets of
risks and chances, which differ not only as to the objects or groups of
objects to be bought, but also as to the point of time at which to effect
the purchase. In doing so he, to some extent, foregoes the opportunity
of better sets offering themselves, to which, after having taken action,
he cannot shift over at all or to which he can shift over only with difficulty
and at a loss. This loss is peculiar in that the danger of it is present even
in otherwise perfectly safe investments. For this the quasi-rent of an
eligible set must provide compensation that will largely account for the
discrepancies under discussion. Making up one's mind, therefore,
involves forming, consciously and subconsciously, a great many anticipa-
tions about the future behavior of relevant factors, the net marginal
result of which, currently corrected, is for the public as a whole theo-
retically measured by those deviations.2 One of these anticipations refers
to the future behavior of interest. Inasmuch as this may be the domi-
nating consideration, we can perhaps say by way of a first approximation
that deviations of the relation between values of investments, similar in
other respects but involving comitments of different duration, from the
relation supplied by our principle of discounting3 express the public's
marginal opinion about the future behavior of the rate of interest.
1 For reasons which, if not obvious, will presently appear, that is often difficult and
sometimes hopeless. Moreover, in highly imperfect situations in which all or some con-
cerns, instead of behaving like drops in a sea, are more like discrete islands conditions
between which are never fully equalized and which have their special markets not only
for their products but also for their factors, it is quite possible that some of those concerns
also have semidetached money markets of their own, partly fed by their own accumula-
tions or the savings of people — officers, workmen — connected with them. In such cases
there may be point in speaking of an "internal" rate of interest as distinct from either the
general one or that of other concerns. Fundamentally, however, and particularly in perfect
conditions, the rate to discount with is always the market rate.
8 As we have seen in Chaps. II and IV, these anticipations may act as disequilibrating
factors, and this as much if they are borne out by events as if they fail to come true. But
this is not necessarily so, and the case in which they have an equilibrating effect is practi-
cally more important than the others.
3 It is easy to see that, even if valuation of durable instruments conformed exactly to
that rule, the values of instruments of unequal durability would be unequally affected by
THE RATE OF INTEREST 611
5. Now, the future balances, which according to our theory are the
price of present balances, do not as such exist, and all the seller of present
balances for the time being receives is promises. These must be embodied
in legal instruments, the types of which vary as much as the types,
situations, and particular purposes of borrowers and lenders — from the
savings and (as far as it is not simply the same thing as a demand deposit)
time deposit, the note and bill, the open account, and so on to the mort-
gage, the bond and, as from our standpoint we have to add, the share.
And since all these embodiments of future balances bear from the actuarial
and the investor's standpoint a surface resemblance to other marketable
sources of returns, much of what has just been said could be repeated
for them. Actually, the same operation is being performed on a sequence
of expected interest and capital payments — or of expected dividend pay-
ments plus expected remainders of equity values — as on a sequence of
expected quasi-rent installments. From this results the difference
between bond rate and bond yield, for example, as well as the relation
between them.1
The fact that titles to future balances differ so much in economic
significance and that even a given economic configuration may be formu-
lated, and the resulting claim safeguarded, in so many different ways,
gives rise to what it is usual to call the Structure of Rates. There is no
objection against thus emphasizing that fact.2 But it should not blind
us to the truth that, as far as this goes, the differences between the con-
stituent individual rates are still only the same kind of thing as the differ-
ences between prices of different qualities of a good or the differences
a change in the rate of interest. Such a change, therefore, tends to alter not only the dis-
tribution of productive resources between transient and durable goods but also the time
structure of the latter taken by themselves.
1 The reader will observe that from that perfectly general standpoint it is no longer
possible to say that quasi-rents will be discounted in such a way that the net return result-
ing from this process will be marginally equal to pure interest plus risk and chance, because
for that purely formal purpose the premium present balances can command bears no other
logical relation to their "capital value" than the one which quasi-rents (though temporary;
this difference disappears in the discounting operation) bear to theirs. What ought to be
said now is that net return of whatever kind — whether interest or quasi-rent — tends
toward equality at the margin with risk and chance (with the addition, perhaps, of such
costs as the act of purchase involves, such as taxes or commissions). But this is only one
of the many relations between interest and other quantities, which we may formulate as
soon as interest pervades a system, and should not be included in any fundamental explana-
tion of its nature. A similar formula equally true in a sense is this: We have seen that
interest is a premium of present over future balances and not, as Boehm-Bawerk has it, of
present over future goods; but insofar as the present balances are intended by "borrowers"
to be spent on present goods, it might be said that interest is a premium of present balances
over present goods. However, this again is more misleading than helpful.
2 There is, however, considerable objection to replacing the rate of older doctrine by an
index of rates, which would obliterate all that is most significant in that structure.
612 BUSINESS CYCLES
between wage rates within an economic domain. If there were no others,
it would still be admissible and for certain purposes necessary to speak
of one rate of pure interest from which the various observable ones differ
by virtue of the element of risk and chance in the above sense; this, in fact,
we have no difficulty in doing in everyday life.
There is, however, a cleavage that is peculiar to interest and that goes
deeper than that, viz., the cleavage between the rates in the three markets
mentioned in Chap. Ill, Sec. E — the Money Markat proper,1 the Open
Market, and the Central Market. It is this which "structures" rates
fundamentally. That the money market and the central market do not
stand on a par and are completing to each other and not competing is
obvious. The open market owes its distinctive characteristic to the
nature of the funds that flow into it and that are being lent, although at
the same time bound to other purposes. The cleavage between this and
the money market is neither in fact nor in logic as sharp as the one
between the latter and the central market. Communication is much
more direct, and money and open markets compete with each other.
The distinction, however, is so important and emphasizes so essential
an element of the modern financial mechanism that we cannot afford to
miss it.
But by thus recognizing the existence in fact as well as in theory of a
fundamental division into three markets of the total volume of trading in
balances, we do not recede from our protest lodged in the last paragraph
of Chap. Ill, Sec. E against another attempt at drawing such a funda-
mental line, viz., between the money and the capital market. We can,
of course, distinguish short- and long-term financing and, for purposes
of description, can think of the latter as concentrated in a special market,
which we may call as we please — capital market or stock exchange,2 for
instance. This special market or rather its component parts are, more-
1 The writer has perhaps to apologize for his clumsy terminology, needlessly departing
from that of financial practice, which uses the term money market synonymously with open
market. It is more important that the above tripartite division does not seem to leave
room for lending by other agencies than banks and quasi-banks. The lending transactions
of households — there is and has been, for instance, in Germany a nonnegligible amount
of mortgage loans given directly by households to other households or firms — and those
transactions of firms which really — and not only apparently, i.e., by means of bank credit
extended to the lending firms — finance open accounts, installments and so on, should be
considered as an appendage to the money market proper, although, of course, they produce
different effects. In many cases firms, especially wholesale firms ("merchants"), must be
looked upon as quasi-banks: in fact they often develop into private banks. We can
thus peremptorily dispose of this matter because it does not greatly affect the drift of our
argument.
2 The stock exchange, which will not come in for discussion until the end of Chap. XIII,
is for us simply an institution for the handling of transactions that belong partly to the
money market and partly to the open market.
THE RATE OF INTEREST 613
over, obvious realities that raise problems peculiar to themselves, and,
in particular, produce their own gross rates of interest in the sense that
is ordinarily meant when one speaks of the structure of rates. In this
sense we shall presently use this distinction, although in doing so we shall
merely speak of different sectors of the money market ; but no significance
deeper than that attaches to it. To say, for instance, that, while balances
are the commodity in the money market, something that is not balances
constitutes the commodity in the capital market (savings or capital goods)
is not only utterly unrealistic but wrong.
Even if it be not held that the money and the capital market, in the
sense of the theories just glanced at, differ in fundamentals, and if it
be correctly recognized that in both it is balances and nothing else that
is being traded in, and that such difference as there is consists only in
the various accessories of the garb in which the promises of future pay-
ments put in their appearance, it is dangerous and often indicative of
faulty analysis to overstress that difference. This easily suggests that
money-market transactions — in this sense — have little to do with long-
term financing and that capital-market transactions never serve the pur-
poses of current business or that the former have only to do with the
effecting of payments (Zahlungskredit) and the latter only with real
investment. It need not be emphasized again how wrong all that is.
Moreover, looking not at purposes but merely at the instruments in which
future balances are embodied, it is not less wrong to think that long
maturities must necessarily be financed by long-term funds. On the con-
trary, it is one of the most characteristic features of the financial side of
capitalist evolution so to "mobilize" all, even the longest, maturities as
to make any commitment to a promise of future balances amenable to
being in turn financed by any sort of funds and especially by funds avail-
able for short time, even overnight, only. This is not mere technique.
This is part of the core of the capitalist process.
The capitalist process develops, along with the money market (in our
sense), perfect negotiability of all instruments of credit, whatever their
legal form may be. Already (Chap. VI, Sec. A) we have insisted on the
importance of the evolution in the course of the fifteenth and sixteenth
centuries of the negotiable note or bill. But long-term financial contracts
such as are embodied in a bond, a mortgage, and also from our standpoint
in a share, became later just as negotiable in principle. In fact, we can
say, if we except a small minority of cases, that for the individual "seller"
of balances no investment in any such title is fixed in the sense that it
binds his money for a protracted or any definite time. Fixity is thus
reduced to, or rather replaced by, the cost incident to actual mobilization
and the risk of not being able to "buy back" the balance without loss.
Of course, a new bond or share may not attain perfect salability until it
614 BUSINESS CYCLES
has acquired standing; its special market may remain very limited forever;
and for these and other reasons its purchase may involve the decision to
keep it for a considerable time or not to sell it at all except when an
occasion offers itself. But this does not alter the principle that in
capitalist society — feudal society, for instance, behaved differently — all
equities and claims can normally be sold at will, irrespective of the purposes
that the sums originally paid in may serve — irrespectively also of the pur-
poses any goods serve that may correspond to them — hence bought by
means of short-term funds. Bonds, for instance, thus become a vehicle
of the shifting of balances, which only technically and by degree differs
from short-term instruments. As soon as this is realized, doubts arise
about the very existence of a distinct thing to be called the long-term
interest rate. The contractual rate on long-time instruments has, of
course, a certain right to that name. But at the moment a contract is
concluded this rate is simply a function of all the conditions of the money
market — barring risks and chances incident to financing the particular
proposition — which are expressed by the short rates. And as soon as the
bonds have come into existence, it is their yield, again a function of short
rates, that counts and not the contractual interest. Braving some danger
of misunderstanding, we may hence go so far as to say that there exists
no such thing as the long-term rate and that, if we nevertheless wish to use
the concept, the thing we ought to mean is some kind of "trend value"1
of short rates.
B. Discussion of Various Rates. — Information about the historical
course of interest exists from very early times, even for the theocracies of
antiquity. About the Graeco-Roman world we know much more. We
know, for instance, what Roman societates publicanorum did to unfor-
tunate provinciates, and have plenty of material about both legislation
and practice. The point that strikes us first of all is the preponderance of
lending for purposes of consumption. Whether protected or enslaved,
the debtor is typically either an aristocratic scapegrace2 or else a poor man
submerged in misfortune. It is this aspect that legislators and jurists
primarily thought of and that they transmitted to the Christian doctors
and canonists — with all the more success because these found exactly
the same aspect exclusively stressed by the Philosophus, i.e., Aristotle.
In other words, they and their medieval successors thought primarily of
1 Lest this should seem to contradict a later statement, it is as well to point out that that
term is here used in a nontechnical sense.
2 In Rome, two distinct types are clearly discernible: politicians, whose career greatly
depended on lavish expenditure on drcenses and other things, G. Julius Caesar being the
outstanding example for the huge element of risk that entered into lending to this type;
and the young man of fashion, whose doings frightened the old gentlemen of the senate
into passing the senatus consultum Macedonianum in order to avoid being killed by sons in
desperate straits.
THE RATE OF INTEREST 615
usury, and our theory lends, in fact, some support to their arguments.
The discovery that debt may open up the road to wealth was not, how-
ever, deferred until modern times. The outstanding case of recognition
of this truth is afforded by a contract that the Romans took from the
Greeks, the foenus nauticum. This provided a method for the financing
of maritime trade mainly by removing or relaxing the restrictions on the
rate of interest in consideration of the clause that the "entrepreneur's"
obligation as to both interest and capital lapsed in case his venture failed
to succeed, i.e.y that he owed only if ship and cargo landed safely. If it be
recalled that we allocate risk to the capitalist, it will be seen how perfectly
this contract expresses the aspect that to us is the essential one.
Data become plentiful from the fourteenth century on, but material
sufficient to derive time series is not available, so far as the writer knows,
before the eighteenth century, during which we can follow the prices of
the English consols. Except for this, serviceable series mostly start in
the second quarter of the nineteenth century, although in important
sections not before the second half.1
1. Success of factual investigation is in this field much interfered with
also by circumstances other than mere insufficiency of material. First,
even if we disregard the problem of the influence upon a given rate of
interest which expectation of a rise or fall in the price level or in the rate
of interest itself may have, it is impossible to isolate the element of risk
and chance. The high rates which ruled in the Middle Ages, for instance,
are largely attributable to the fact that, particularly in loans to princes,
the lender had to consider the likelihood that he would not be repaid at all.
It is impossible to tell how much would have remained of the 80 per cent
that Frederick the Fair, the gallant king of Germany and duke of Austria
(fourteenth century), is reported to have "paid," had the financial habits
of that prince been more regular.2 Cases of this kind abound at all times,
1 See, however, the series beginning September 1795 of prices of 60-day bills on London
at Boston and New York in Smith and Cole, Fluctuations, App. E. Their discount series
(ibid., Table 74) starts in 1831. The best (monthly) American series, the interest rate on
60 to 90-day commercial paper in New York, is in the writer's opinion the best in any
country. It has been used by Professors Crum and Frickey, and the writer's work has been
primarily based on it. For Germany, see Kahn, Die Geschichte des Zinsf usses in Deutsch-
land seit 1815 (1894); Homburger, Die Entwicklung des Zinsf usses in Deutschland, 1870-
1903 (1905); Albrecht, Die Geschichtliche Entwicklung des Zinsfusses in Deutschland,
1895-1908 (1910); and finally, Voye, Ueber die Hohe der verschiedenen Zinsarten (1902),
who for Prussia covers the period from 1807 to 1900. Together, these authors tell a pretty
complete story.
2 This, by the way, should be borne in mind by historians who report on the sorry facts
of usury. To the interest rates recorded correspond the records of the failures of financial
houses that did that sort of business and were in many, if not most, cases bankrupted by it.
On the other hand, we have seen that lenders reaped other advantages besides interest
(Chap, VI).
616 BUSINESS CYCLES
and so do cases in which the lender, by hiding the actual interest behind
some supplementary charge or concession, avoids quoting a figure that
would shock public opinion. All this has an important consequence for
the behavior of some of our series, which, moreover, change their char-
acter as time goes on. Ever since the English government, for instance,
began to enjoy that confidence which it justified for so long a time, the
price of consols varied much as we should expect in a safe — although in
our sense never riskless1 — annuity. But this was not so for the greater
part of the eighteenth century, when we find consols varying, though not
so strongly, in the same direction as stocks did. Similarly, even the
highest class of American railroad and industrial bonds display the influ-
ence, in sections of our period, of a decreasing element of risk.
Moreover, regional differences are sometimes so important as to make
it difficult to speak, even for one and the same type of loan, of the rate
of interest in a country, while interregional and international capital
movements interfere with the behavior of interest rates in any particular
country. Again, taxation, fear of taxation, and the general attitude of
public administration toward interest will all have effects which are not
easy to discern, because they may assert themselves with lags and in
very many forms, differently also for different kinds of titles. In cases in
which placing of bonds is difficult and costly, or where intervention on
the market is provided for, there is an element of remuneration for these
services. Some rates are very slow in responding to changes of the situa-
tion. Others mean different things or have varying importance not only
in different countries but also in the same country at different epochs.
The American commercial paper rate is a conspicuous example.
Even if bankers' banks did no business except with member banks,
it would not follow that there is only one rate in the Central Market.
A central bank is a discriminating monopolist, although not a monopolist
that aims at maximizing net revenue. For many purposes this can be
neglected, however. Similarly, the Open Market evolves several rates
which at times differ significantly but still present, normally at least,
a picture so uniform as to warrant speaking of the open-market rate.
The money market in our sense, however, splits into so many sectors and
subsectors that no theoretical conviction to the effect that there is,
after all, some meaning to the concept of a single rate of pure interest
also in this market, avails against the fact that it is impossible to indicate
it in a satisfactory manner. As we have seen above, we may schematize
those sectors and subsectors according to types of borrowers — the German
peasant, for example, had hardly any sources of balances to turn to other
than saving banks, cooperative banks (Genossenschaftskassen of the
1 The late Alfred Beit, the South African financier, used to say that he never lost so
much on anything a§ he did pn consols,
THE RATE OF INTEREST 617
Raiffeisen kind), and village usurers; to German governments and
municipalities any source was open — according to types of funds — certain
household as well as institutional funds were even by law directed into
certain channels; and large classes of savers habitually turn to certain
kinds of investments and not to others — and according to types of title —
from the standpoint of the issuing corporation, as from the standpoint of
theory, stocks and bonds are fundamentally the same thing and yet,
owing to the difference in the rights they embody, they suit different
people and different situations. Markets of a higher order of specializa-
tion, in fact, develop around every issuing house or financial group as
well as around every industry.
In briefly surveying some of the more important of these special
markets and the interaction of the rates that rule in each of them, we
will discard the market (which again splits into many subsections) of
loans to households, except loans to "public households.'* As has been
already mentioned, however, an element of lending to private households
for purposes of consumption is included in agricultural credit. Also,
we will not specially deal with that mass of credit that arises from the
relations between manufacturers, wholesalers, retailers, and households,
though not all of it is financed by banks. And we shall finally neglect
the problem of regional differences.
2. Many authors, notably German ones, would look for the rate of
interest ruling in a country (landesublicher Zinsfuss: what a suggestion
of traditionality, ominous for our purpose, this turn of phrase conveys!)
to the interest charged on rural and urban mortgage loans. There can,
however, be an approximation to truth in this view only in the case of
first mortgages on urban realty. Other than first mortgages on urban
realty are in many cases pure speculation. As to other than urban
realty we must bear in mind, on the one hand, that in many countries
habits of certain classes of savers, as well as legislation regulating invest-
ment of trustees' funds and investment by saving banks, favor this kind
of credit, while, on the other hand, mortgages on farm land imply a
special kind of risk, very deterrent to many people, i.e., the risk of having
to undertake the management of the farm or to find a manager for it.
These facts go far toward setting up a special market displaying a special
rate. In any case, we cannot look for a true representation of the cyclical
behavior of the rate to those mortgage rates that are very sluggish from
year to year, except in the case of the Kondratieff.
Of course, though some reservations are necessary owing to the organ-
ization1 of this kind of credit, these rates do eventually follow the general
1 This organization perfected to the utmost in Germany during the last 20 or 30 years
of our period and even before that highly efficient, accounts, for example, for the fact that
the rise in interest that begins in 1895 asserted itself but little in this market.
618 BUSINESS CYCLES
march of things. There are, moreover, two major links with other sec-
tors of the money market. On the one hand, only part of the total
amount of mortgage credit consists in balances that the lender intends to
tie up for a long time. Other sources are available and, at some times and
in certain countries, particularly in the United States, banks have gone
much further in extending credit in those directions than principles of
classic banking practice — and common sense — warrant, so that mortgage
money coming from nonbanking sources often competes with the ordinary
sources of credit provided by commercial banks.1 On the other hand,
mortgages may be "mobilized" like every other instrument of credit,
and then they invade the bond market and establish connection with
the latter's rate. The most perfect instance of this is the Pfandbrief of
the German mortgage banks (Hypothekenbanken) or of cooperative credit
institutions of the type of the Landschaften. The actual rate of interest
for this kind of credit was not simply the rate which the borrower con-
tracted to pay but that rate taken with reference to the price at which
the bonds sold in the market2 and also to the charge made by the insti-
tutions for their services. We have various series, especially since 1870,3
which very clearly bring out the fact that at this point mortgage credit of
the agricultural and urban kind merges into the general bond market.
In view of what is to follow later, we may note that interest on "true"
mortgages — distinguished from mortgages that only secure other types
of credit — as reflected in the rate and yield of Pfandbriefe (the writer
prefers not to translate the term) fell from 1870-1895 and then rose till
1914, although not very much. The Preussische Central-Bodenkredit-
Aktiengesellschaft, for example, which offered its Pfandbriefe publicly,
issued in the year of its foundation (1870) a 5 per cent type at par (which
means 5^ per cent for the borrower). In 1879 it financed on the basis of
4-Hz per cent at par, in 1884 at 4 per cent, in 1889 at nearly 3j-£ per cent.
From 1890 to 1894 it had to return to the 4 per cent type. It reached the
minimum in 1895, after which a slow and moderate rise set in (see Horn-
burger, table on page 97).
3. The bond market is another semiindependent piece of the mecha-
nism which we call the money market. The fences that, worn away in
1 The reader will, however, recall that banks, when they lend for agricultural purposes
other than current business or for urban building, do not always commit that mortal sin
against the classic rules of banking which has so much to do with difficulties in depression.
Besides, it has been mentioned already that mortgages may be no more than an additional
security for an unexceptionable credit.
2 This calculation presents various difficulties into which we cannot enter here. In
particular, it would, in case the bonds are above par, be correct only if the whole premium
went to the borrower, which in practice was not generally the case.
3 See Hecht, Organisation des Bodenkredits in Deutschland, statistics of mortgage
banking in Part II, vol. 1 ; also Homburger, cited before, p. 68, et seq.; for older data, see
Von Cyriaci-Wantrup, Agrarkrisen und Stock ungsspannen, p. 87.
THE RATE OF INTEREST 619
many places but still recognizable, separate it from other sections consist
not only of costs and risks but also of habits of investors and positions of
borrowers. Their reality is proved by the fact that, in the short run,
bond yields sometimes differ considerably not only from "short" but
also from other "long" rates. But the bond market is also very liable
to split into subsections — so much so that different issues of the same
borrower occasionally display an individualism (see, for example, the
different behavior during the 20 years before the war of the German
government 3 and 3% per cents) that is not always easy to explain.
One important subsection the individuality of which is accentuated
by legal privileges of various kinds and by a strong, if sometimes uncrit-
ical, preference on the part of savers in some countries — particularly in
France — is constituted by government bonds. In England and Ger-
many, however, this market received a shock by the policy of conversions,
which in part at least sprang from an erroneous diagnosis of the nature of
the general fall in rates to 1897 and possibly also from an overestimation
of immediate gain as against future disadvantage. In the United States,
confidence in the currency established in 1879 or, practically, in 1878, was
repeatedly shaken by soft-money agitation and free-silver campaigns,
one of which is perhaps responsible for the small peak in yields that
occurred in 1896, when the yield of the United States 4 per cents increased
by J4 per cent. But the policy entered upon in the late eighties of buying
up, regardless of cost, influenced quotations much more strongly in the
opposite direction. The fact that the yield of United States 4 per cents
fell till 1901, unlike that of German governments which touched the low
point in 1894 and that of English consols, which did so in 1897, * may
possibly be explained by this, although other bonds did the same thing.2
1 The high point in the London and Cambridge Economic Service Index of the prices
of fixed interest securities, which, however, gives rise to various doubts, comes in 1896;
the French rente reached its maximum in 1897.
2 Mr. Macaulay's Index of American railroad bond yields (the " yield of best bonds each
January") touches, however, its low point in 1899 (Journal of the American Statistical
Association for March 1926. The reader should refer to this paper for questions of tech-
nique and some points of, comparatively speaking, detail). Moody *s monthly series of
corporation bond prices, 1866-1914, described in Moody's Investors' Service, November
1924, reaches a high, and the Standard Statistics Index of Yields upon 60 High Grade Bonds
a low point, as late as 1902. The Treasury's policy of debt redemption beyond sinking
fund requirements, so vigorously pursued, may in part account also for this, because a
considerable fraction of the funds set free thereby would understandably seek investments
similar in character. The movement, contrary to expectation, lasts, however, exactly as
long as the sharp increase of investments by National Banks. Besides, debtor concerns
certainly became better as a rule after the crisis of 1898 was over. There is no need,
therefore, for concluding that, as the yield of governments behaved substantially as other
bonds, we are faced with a true movement of "the long rate" asserting its independent
existence and indicative of deep-seated causes in the sphere of "real capital."
620 BUSINESS CYCLES
All this need not detain us. The general contour stands out clearly.
As to other than government bonds, the reader should, however, bear in
mind that in this country the market of prewar days differed fundamen-
tally from that of the twenties of this century in material, in public, and
in financial structure. Railroad bonds were the big item. They and,
still more, the bonds that the merger movement produced were primarily
taken in large lots by institutional and wealthy private investors. The
general public had not much to do with them, and Europe gradually sold
out. In England, also, railway debentures were the most prominent
"business" element of the bond market, which dealt mainly in local,
municipal and quasi-municipal, and foreign loans. In Germany a market
in industrial as distinguished from railway bonds, important enough to
figure and to display regularities of its own, developed in the course of the
eighties. Many factors, international influences, and changing quality
or security or changing financial standing of debtors, among others,
would have to be taken into account in order to explain the behavior of
rates in this market and the changing relations between the rates ruling
in its various sectors. For us, however, it is sufficient to complement our
statements at the end of the preceding section with a few remarks on the
mechanism which links bond yield to short money rates.
In prewar times, and particularly in London, speculators of a certain
type — they were really a kind of arbitrageur — reacted promptly to very
small variations in the relation between bond yields and short rates, and
at once borrowed short in order to buy bonds when the margin covered
cost and risk. In this country the negative short-time association
between the loans and the investments of banks outside New York City,
to be mentioned again in the next chapter, is indicative of another link.
And the financing of subscriptions to bond issues by bank loans and the
refunding of bank loans from the proceeds of bond issues provides a
third. We readily see not only that this mechanism is by nature inca-
pable of equalizing bond yield and short rates, owing to the particular
kind of commitment incident to the legal construction of the claim
embodied in a bond, but also that for reasons of financial technique it
will be incapable of enforcing instantaneous, still less proportional,
covariation. Even cases of variations in the opposite direction are by no
means rare. For instances we need not look only to situations of deep
depression, in which, when the panic is over, short rates may drop almost
to vanishing point, while bond rates and bond yields stay up. It can
happen in the most normal course of things that, for example, when a
large issue has been negotiated and the borrower next lends the proceeds
in the open market, bond yields rise and short rates fall for the time being.
International relations, though we do not now take them into account,
may be more effective in the one market than they are in the other. A
THE RATE OF INTEREST 621
sufficiently widespread tendency among industrial borrowers to make
themselves independent of their banks may occasionally raise the one and
depress the other rate. It also should be clear, however, in what sense
we may say that all this is "merely technical." In spite of it, the bond
market is fundamentally but the most sluggish part of the money market,
and in its rates reflects, though sometimes with a lag and always at a
higher level and with smaller amplitudes, the fluctuations in short rates,
which not only express business situations more faithfully because they
are free to react instantaneously, but also are the main influence in
shaping bond yields. The behavior of the two rates relatively to each
other, in fact, bears out — although, by itself, it does not prove — the
consequences that follow from this : it has for this country been shown by
Professor W. M. Persons that the graph of bond yield, plus one-half of
one per cent, looks strikingly like a "trend line" drawn through the graph
of short rates.1 Deviations from this evidently basic relation are dis-
tinctly secondary and can easily be accounted for.2
The problem of stock prices will be dealt with in the next chapter.
Here it is sufficient to repeat that, according to the view taken in this
book of the shareholders' position and of the nature of their dividends,
the stock market is simply another section of the general market of bal-
ances. Since profit is for us the pillar of interest, dividends per cent of
capital actually paid in, which bear a particularly close relation to profits,
are really the most fundamental of all the rates of the money market.
Broadly and on the average, they move in the same direction as bond
yields (or rates on new bonds) and short rates. But the dominant
influence of those risks and chances that are peculiar to common and,
to a lesser degree, to all stock, impairs the working of the mechanism that
1 Compare his study in the Review of Economic Statistics for April 1927, chart on pp. 94
and 95, see Chap. V, Sec. B, where this was given as an instance for a "reference trend."
Crossing points roughly coincide with the inflection points of a smoothed curve of short
rates and are — again, roughly — indicative of the Kitchin cycles. Otherwise they are not
easy to interpret. There is, in particular, "not always a one-to-one correspondence
between crossing points, on the one hand, and peaks and troughs (of the curves for four
series of security prices) on the other hand"; nor is there any "pronounced tendency
revealed for crossing points, on the one hand, and peaks and troughs, on the other hand,
to occur simultaneously" (op. cit., p. 101).
2 See Dr. C. E. Thomas's paper, The Effect of the Depression upon Bond Yields;
Journal of the American Statistical Association, September 1988. His result (though arrived
at mainly from postwar material) that in an approximately normal situation a decrease
of 1 per cent in the short rate would produce a decline of approximately 0.24 per cent in
bond yields (p. 267) is as suggestive as is his emphasis on the influence of trade volume, any
decline of which tends to counteract the influence of a fall in short rates and to reduce
it practically to zero precisely in "deep" depression (p. 271). It is important to bear in
mind that this does not contradict any of the statements in our text. Volume of Trade
influences revenue, revenue influences the safety of the bondholder's investment (op. cit.,
p. 262).
BUSINESS CYCLES
links this section of the market to the other two. Neither investors nor
"borrowers" can choose between purchasing and issuing bonds or stocks
according to actuarial rationality alone. And there is no simple and
reliable relation between money rates and stock prices such as there is —
barring risk of default — between money rates and bond prices. The
primarily negative association between stock and (highest grade) bond
prices, however, not only indicates that that mechanism is after all not
absent, but also reveals the way in which the counteracting influences of
variations in returns (dividends) and variations in pure interest work out
in the course of cyclical situations.1
4. In one way or another, all the types of transactions in balances
that we have so far discussed thus lead toward the Open Market. This
is contrary to expectation, inasmuch as our schema has on its monetary
side been made to center in the relation between member banks and their
customers. We would, in fact, select for guide in our analysis the interest
charged to customers in current account (Kontokorrent-Kredit, cus-
tomer's line of credit), the true bank loan rate, if we could. From this,
however, we are not only estopped by the impossibility of compiling a
reliable series for prewar times, but also by those difficulties inherent in
this material of which mention has been made above. This market is so
far from perfect, charges differ so much not only locally but also as
between banks,2 customers and transactions, and are so sluggish, that
tendencies can be read off much better from the series of the New York
commercial paper rate3 or, for England and Germany, from other open-
market rates which in those countries more or less correspond to that
American series. This choice is not satisfactory. The open market
and its immediate vicinity, indeed, enjoy greater (almost absolute)
freedom from those influences that make for stickiness and are almost
perfectly competitive. But they, more than other sectors, are exposed
to the regulative activity of central banks — degree and method of which
greatly varied during the long period under discussion — and to the dis-
ruptive influence of the factor which, especially in England, it was the
principal aim of that activity to control. This factor is the mass of
1 Cf. W. M. Persons, Money Rates and Security Prices, Review of Economic Statistics,
January 1926.
2 In some countries, particularly in Germany, this situation produced the result that
is so often the outcome of market imperfections: banks entered into agreements with each
other (Konditionenkartelle) with a view to normalize charges. But this affects the last
prewar decade only.
8 To the published figure between ^ and 1 per cent should be added, in order to arrive
at the rate actually paid by borrowers. The remuneration for the series of intermediate
dealers is, in general, not comprised in published open-market rates. The relation between
the rate on 60 to 90-day paper and 4 to 6 month paper is as we should expect, and does not
call for comment.
THE RATE OF INTEREST 623
temporarily unemployed balances and credit facilities, which in that
market hunted "for any wretched rag of interest."
It is here that what we have called Temporary Investment puts in
appearance. Funds which large concerns have accumulated for later acts
of real investment or simply as a cash reserve or which they have on their
hands when they are underspending in depression, or surplus funds of
public or semipublic bodies, such as the Prussian state railroads and the
Prussian government or India House, flowed, with increasing rationali-
zation in the management of funds, increasingly into the open market for
normally riskless and ideally short investment. In these cases, banks
were frequently employed as agents. But they did a similar thing on
their own account if their own credit facilities were underutilized in their
own business and also in order to collect their secondary reserve (see
Chap. XIII). They even, especially in England, shifted to the open
market part of what otherwise would have been their normal business
with their own customers : bill brokers (discount houses) looked after the
acceptors of that part of their portfolio and relieved them of much trouble;
this business was impersonal and restriction was normally possible at any
moment without loss of either money or prestige and without remon-
strances from irate customers; while firms such as the merchants of
London and the great provincial centers were glad to offer their remit-
tances in Lombard Street, seeing that as a rule they got lower rates from
their broker than they would have from the very bank that financed that
broker. This means, of course, that the amount of actual and potential
balances offered in that market, while very variable, is extremely inelastic.
No fall in rates discourages it; save in exceptional circumstances, such as
obtain during speculative manias, it hardly increases in response to a rise
except by way of immigration of foreign balances.1 This, together with
the perfect competition which rules in the open market, accounts for the
frequency and the amplitude of variations in its rates, as well as for the
close covariation of all of them.
All this undoubtedly deprives these variations of much of the signifi-
cance they would otherwise have. It should be remembered, however,
that this significance is not further reduced by the short-time character of
the operations which those funds serve. For we know that this character
is largely delusive : at least through financing the carrying of stock, even
call loans are by no means debarred from financing such things as railroad
building, while other forms of open-market credit much more directly
serve the purpose of long-time real investment. Finally, we have to add
to our list of reasons why rates of interest should lag behind other ele-
1 More correctly and also more generally : by way of immigration into the banking circle
under consideration; for large countries may so divide up into sectors that a rising rate
may be effective in drawing floating funds from one to another.
624 BUSINESS CYCLES
ments in the cyclical process one more item that is peculiar to open-market
rates, which are so remarkably free from the influence of others. As far
as the balances floating in the open market have their source in under-
spending, they obviously cannot act on rates until restriction of indus-
trial and commercial operations has occurred to a significant extent, that
is to say, when liquidation has already turned into depression. Then
the effect of decreasing borrowing is intensified by the increase in funds
offered. The causal importance of this for the puls"e of business is,
however, small, as we shall presently see. And the location in time of the
upper turning point in the money curve of the Harvard Index Chart is not
as a rule explained by it. But the lag of the lower turning point in the
interest series may be, for until the slack due to underspending is taken
up, open-market rates cannot recover from their trough.
Call (broker's loans' renewal) rate and time rate do not differ more
than we should expect from the smoothing effect the time arrangement
must have. Commercial paper rate varies much as time rate, though at a
somewhat higher level. No systematic change in the relation between
the two seems to have occurred in this country or in Germany during the
last 30 years before the war and not much in England, but there is in it
a characteristic cyclical variation. We will, however, confine ourselves
to a few comments on some of the instruments of credit that are mainly
financed by open-market balances.
Treasury bills — we use this term for all perfectly negotiable short
government obligations and for all countries — are, from the standpoint of
the buyer, merely the senior members of the family of bankers' acceptances
and move with the rest at a premium, which understandably widens in
prosperities and shrinks in recessions. They may, for our purposes at
least, be said to mean much the same thing in our three countries.1 But
bankers' acceptances, and therefore their rates, do not. They are first
the typical instrument for the financing of current commodity trans-
actions in international trade, and as such they have, up to the end of the
seventies, primarily belonged to the English domain. It was not until
the eighties that Germany increasingly took to financing her foreign
trade herself. Something of that close relation to individual commodity
transactions, at first so characteristic for this instrument of credit, was
lost, even in England, in the case of the finance bills that used to serve,
and still serve, in American export trade. But English practice sub-
stantially confined itself to this particular purpose and did not knowingly
use finance bills for any other to a significant extent.
In America bankers' acceptances of this type, and a market for them,
are a war and postwar growth. Before the war, accepting was even ultra
1 The Austrian Salinenscheine afford the first European instance known to the writer
of open-market operations undertaken by a government for purposes of regulation.
THE RATE OF INTEREST 625
vires for National Banks. But in Germany bankers' acceptances were
also used to a very large extent (much diminished after the war) for
purposes of domestic trade and not only for that but even as a means of
raising funds for real investment and of financing plant and equipment.
Thus, the bankers' acceptance, which costs the bank nothing but its
signature, became so popular as to acquire a central position in the market.
A technique grew up for trading in these acceptances which made them
ideally salable at a single rate by standardizing the "commodity" through
certain requirements as to the standing of the accepting bank and as to
form (round figures; observe the characteristically different attitude of
English banks in this point). They were even directly used as means of
payment. There is no perfect parallel to this in England or America,
but the writer has come to the conclusion, in spite of considerable doubts,
that this rate (Privat-Diskont) is the nearest equivalent to the New York
commercial paper and the Lombard Street rates. The decreasing impor-
tance of the commercial bill must, however, be borne in mind: with
English banks, for example, bills were 26 per cent of assets in 1880 and
only 1£ per cent in 1912.
5. Finally, it must be borne in mind that although in our schema
bankers' banks and their rates stand "at one remove" from the money
market, this was not so historically. Not until the Federal Reserve Act
did any bankers' bank confine itself to bankers' bank business only.
On the contrary, they all did member banks' business also — at times,
primarily — and their central positions and bankers' bank policies grad-
ually evolved from it. In the case of the New York banks, to which,
nevertheless, we feel justified in attributing central bank functions, this
precludes any search for a bank rate in the sense of our analysis. The
Bank of England's rate was, in the times of the bill and the bank note,
simply the rate at which it was prepared to do business precisely with
nonbanking customers. This rate was highly inflexible (between 4 and
5 per cent) for over a century, until Peel's Act (except in the latter part of
1838). Then what was called the New System of Discounting was
embarked upon. It consisted in competing for paper in the market and
resulted for a time in a dominating position of the Bank in this business.
Rate policy was currently adapted to that purpose, and although it would
not be correct to say that the Bank merely followed the market — a buyer
who acquires 50 per cent of any commodity can never be said merely to
"follow," (see the next chapter) — its rate was certainly just one of the
market rates.
Up to 1858 the Bank at least rediscounted for bill brokers and dis-
count houses, thus doing typical bankers' bank business, which practi-
cally amounted to rediscounting for banks in the usual sense of the word1
1 Bill brokers and discount houses, it should be recalled, are for us a special type of
bankers or "satellites" of bankers. What is meant above is that the London banks which
626 BUSINESS CYCLES
and in effect overcame the It Is Not Done which obstructed the march
of the latter to the Bank. But in that year rediscounting for the bill
market was discontinued, and the Bank thenceforth entirely relied on
its ability to attract the desired amount of paper competitively1 until the
early seventies, when this practice ceased to give satisfaction to the
public. Until then bank rate, though as a rule on a higher level, would
serve our purpose almost as well as market rate. But after some years of
hesitation, the Bank then gave the strongest proof of its determination
to cling to its member bank business. In 1878 it declared that it would
no longer adhere to its official rate when discounting for its own cus-
tomers but charge them about the market rate. At the same time, it
made a move to retrace its steps in the matter of rediscounts by inti-
mating its willingness to make special advances to bill brokers (discount
houses), although rediscounting for them — at the official rate — was not
resumed until 1890.2 These two measures are highly significant precisely
because they were taken together. They mark a new stage in the evolu-
tion of the Bank's policy: while it asserted its member interest, it simul-
taneously divorced this from its central bank function; bank rate ceased
to be the hybrid that it had been until then, and began to assume the
place assigned to it in our schema.
Similarly, the rate of the Prussian and later, the German Imperial
Bank was primarily a discount and only secondarily a rediscount rate.
Although the Bank was always under pressure from popular demands for
a cheap and especially steady rate and at times had to yield to them, we
may repeat for Germany, what has been said in the case of the Bank of
England, that, for the nineteenth century at least, we could do almost
as well with bank rate as with any market rate.
C. Discussion of the Time Shape of Interest Rate. — This discussion
will primarily be based on the series presented in the pulse charts, which
also show how interest behaves relative to other elements of the cyclical
process (see also Chart XX) . Call rate (Charts XXXVI and XXXVII),
in its relation to stock exchange figures, will be briefly discussed in the
financed brokers were enabled to call their loans at will, since the latter could go to the
Bank, and that this put them in much the same position as if they had discounted and then
rediscounted themselves.
1 Under conditions of imperfect competition that is, especially in England, entirely
compatible with being dearer than one's competitors. In its provincial branches, however,
the Bank discounted at the prevailing local rate and sometimes below it.
2 Mr. King in History of the London Discount Market, p. 304, stresses the later rather
than the earlier date and thus misses the essential complementarity of those two measures.
He even (p. 297 n.) accuses Palgrave of inaccuracy in holding that the policy of 1858 was
reversed in 1878. But Palgrave does not seem to assert much more than Mr. King admits.
The difference seems to arise from an inclination of Mr. King to underline the achievements
of William Lidderdale very strongly. His own interpretation of the resumption of redis-
t, however, impaired thereby.
THE RATE OF INTEKEST
next chapter. From the outset it must be borne in mind that the rates
chosen, although comparatively "pure," still display the influences of
legislative changes, currency and credit policies, gold production, inter-
national relations, and in particular of panics — all of which unavoid-
ably interfere with expected behavior and with such barometric value as
the rate of interest may be held to have. But a barometer that can
be spoiled is still a barometer, and in large intervals within our period
some of these influences — though not that of panics, which acted on
interest much more strongly than they did after the World War — were
not very important.
1. Since we do not use indices of rates, interest series are natural. Of
course, they are systematic — in fact, they may justly be called the most
systematic of all. There cannot be any doubt that what they describe
is a primary element of the cyclical process of evolution, though not all
economists will agree to the statement that follows from our analysis, viz.,
that interest is fundamentally consequential, and is causal only in a
secondary sense. Deferring additional comment on the latter point, we
note that, similarly, everybody will accept the proposition that interest
is the most cyclical of all the elements of the system; but hardly anybody
the further proposition that it makes a "clean" cyclical series. Improve-
ment of the organization of credit, decreasing premia for risk, and other
such circumstances may account for a falling descriptive trend in interest
series, but our theory denies that there is any result trend. While in the
pure model interest would in each cycle start from and return to zero,
there is no generally valid reason why it should not in actual fact come as
near to it in any neighborhood of equilibrium — account being taken of
the phases of underlying cycles — as it does in any other.1
Data good enough to test that proposition under comparable condi-
tions over a sufficiently extended period do not and cannot exist. But
such information as we have about centers of capitalist evolution anterior
to our period is corroboratory rather than otherwise. In Amsterdam,
market rate in the eighteenth century — even in the seventeenth — was
sometimes as low as 2 per cent.2 However hazardous the comparison —
1 There might be a special reason, however, incident to the factor of consumers' borrow-
ing: the influence of bourgeois mentality which had created a parsimonious state, being
displaced by the influence of another mentality, making for an extravagant state, may
produce a noncyclical rise in interest in the future or, as an alternative, currency disorders.
Such conditions actually did prevail before the reign of the bourgeoisie and so did the conse-
quence envisaged, while the opposite consequence actually ensued as those conditions
passed. There may also be spurious reasons: as pointed out in the text, modern technique
does not allow interest to soar to panic peaks.
2 Even in the England of Walpole and Pelham, 8 per cent consols — the price of which
we could not, however, for reasons stated in Sec. B, take as a reliable guide — occasionally
sold at prices higher than any they ever reached afterward (maximum, 107), see W. St.
Jevons's chart in his volume on Currency and Finance.
BUSINESS CYCLES
in actual fact, minima cannot always be taken to indicate neighborhoods
of equilibrium, see infra — it is not without significance that in 1897
London market rate was about as high (average for that year, 1 pound,
18 shillings, 6 pence). And again the question seems in order as to how
much the reader thinks would have been left, in either case, of those 2 per
cent if there had been no consumers' borrowing. Moreover, inspecting
the graph of our series of New York commercial paper rates from 1875
to 1914, we find that it fluctuates about what in a first approximation
seems to be a perfectly even level. In fact, it looks as if a trend had been
taken out of it which may, however short the series, be adduced in favor
of our thesis, because it is obviously not likely that if a systematic tend-
ency to fall were present in interest rates, this tendency would not have
shown at all within 40 years of both vigorous and relatively undis-
turbed capitalist evolution.1 It should be observed in passing that these
facts, by refusing support to the so-called Law of the Declining Rate of
Interest or, as the classics said, Profits, indirectly also cast doubt on the
theories of which that law is a consequence. Nor should we be surprised
at this. There is nothing in theory or fact to justify belief in any tend-
ency in the recurrent waves of profit (in our sense) either to increase or to
decrease systematically. It follows that interest fundamentally deriving
from profit should also not display such a tendency.
2. The reader will readily see that expectation as to the cyclical
behavior of interest, if we adapt it to the higher approximations of our
model, including the coexistence of three cyclical movements, will be for a
lagged rise in prosperity and a similarly lagged fall in recession.2 Since
the neighborhood value of interest rate is not actually zero, further fall
will, in general, occur in depression as a consequence of abnormal restric-
tion of volume of business. But as in all other cases, it must be remem-
bered that the processes of depression are erratic, even apart from the
effects of panics on the rate of interest which are peculiar to it. Uncer-
tainty extends, in consequence, to its behavior in revival, which should
bring it back to neighborhood value. Again it must be emphasized that
these expectations will work out somewhat differently in cycles of different
span.
1 Mr. Carl Snyder, Interest Rate and Business Cycle, American Economic Review,
December 1925, p. 697, notes that "it is striking how often this peak [of cyclical expansion
as measured by the Clearings Index of the Federal Reserve Bank of New York] has come
close to the time at which the interest rate has risen to just above the average of the half
century, i.e., when the interest rate crosses the average line" (which is at 4.98 per cent, 1875
to 1925). This, again, could not be if there had been a significant trend.
2 The writer has sometimes heard and read the statement that the fluctuations of credit
lag behind fluctuations in business activity. This is not the same thing — and not quite
correct — although it is obvious that the statement means the same thing, i.e., that interest
lags: other elements of the sphere of money and credit do not.
THE RATE OF INTEREST
On the whole, this is what we find. Owing to the abnormal sensitive-
ness of our series and to the fact that short money rates reflect, and are
connected with, everything that happens in the business world from day
to day, surface movements are much more, and underlying movements
correspondingly less, accentuated than in a series of employment or
pig-iron consumption. We therefore find that the graph, while it dis-
plays many other surface movements also, is dominated by the Kitchin
wave, presence of which is obvious at the first glance. Interest rate is,
in fact, the standard instance by which to demonstrate this cycle and has
been repeatedly analyzed with a view to measuring it. For us it is
sufficient to refer again to Professor Crum's periodogram analysis of the
New York commercial paper rate series (Review of Economic Statistics,
1923) which has been several times quoted in this book, and to point out
that the phenomenon can readily be seen also in the English and the
German series. Nevertheless, as has been emphasized by Professor Crum
and as is proved by the fact alone that different authors have invariably
derived somewhat different periods, the results of formal analysis are not,
in themselves, entirely convincing. They cannot be expected to be so
with material that is subject to so many internal and external irregulari-
ties (see Chap. IV, Sec. D). No one with the slightest familiarity with
business responses can ever hope to "prove" this or any other cycle from
time-series evidence alone, since this would require that periods, ampli-
tudes, and behavior in phases be entirely unaffected by any external fac-
tors and always the same in war and in peace, in times of increasing and in
times of decreasing gold production, in times of sound money and in times
of actual or threatened disorders. Still, such short and such strong
fluctuations, which also are, comparatively speaking, so regular, afford an
exceptionally favorable case. We cannot hope for equally good results
with regard to any much longer ones. Formal methods would hardly
yield results in the case of the Juglar, and if formal measurement be
insisted on, it is easy to deny its existence. But if the reader follows our
series year by year, starting with the characteristic rise in 1879, he will
be able to mark off the last two Juglars of the second, and the first two
of the third Kondratieff. Some well-marked troughs are a help — e.g.,
1885, 1894, and 1904 (1908 requires careful interpretation) — but those
peaks which are immediately followed by abrupt fall of over 3 per cent
should be neglected. Of course, in this case, as in all others, conviction
can be attained only through a knowledge of what actually happened,
strictly speaking from year to year, in the industrial organism. Chapters
VI and VII seem, however, to provide all that is necessary for a first
attempt.
In the series of commercial paper rates and its equivalents in England
and Germany, the Kondratieff should show in what traditional analysis
630 BUSINESS CYCLES
would describe as breaks in trend — as it does in other series that cover
about the same stretch of time. Stating the question is answering it.
Everyone knows that rates displayed a long-time tendency to fall until
the middle nineties, when an opposite tendency began to assert itself,
although in the United States — not so, as we have seen, in Germany
and England — bond yield failed to follow suit for some years. It is
possible to draw a plausible declining "trend" through the New York
figures from the beginning of the series to 1897 — the minimum occurs in
1894 — and a not less plausible rising one through the figures for 1897 to
1913. The same can be done for England — where the minimum in mar-
ket rate occurs in 1895 — and Germany. This means that we have only
a slight and very short recovery rise, which, however, is for the same
reason as in the case of price level not astonishing in a Kondratieff, the
long and gentle sweep of which is little influenced by spirals. We may
go farther back, although on decreasingly satisfactory material. The
rise in yields of English consols toward the end of the eighteenth century
does not, as the rise in price level, precede the French wars, but coincides
with their beginning. The all-time peak of 1798 could be sufficiently
explained by those wars and inflation alone. The wars also explain the
fall of Prussian governments, which yielded over 16 per cent as late as
1813. This casts doubt on such evidence as we have about interest rates
from 1787 to 1815 at least, but after the latter year we are entitled to
speak of a tendency of interest to fall for a period much too long to be
accounted for merely by the process of absorption of war effects, some of
which moreover would have tended to raise rates.
That period lasts in England and Germany to between 1842 and
1845. For this country disorders of the wildcat banking type and cheap-
money policies explain violent upward spurts in the thirties (1834, 1836 to
the beginning of 1842; see Cole and Smith, op. cit., table on pp. 192 and
193) which blur the picture: rise indicative of the processes of a Kondra-
tieff prosperity here begins in 1845. So it seems to have been also in
Germany, and this is a more serious deviation from expectation, since it
cannot be accounted for by equally significant disorders in the sphere of
credit during the late thirties. For 1844 (see Kahn, op. cit., and Voye,
op. cit.) bond yield, mortgage interest, and the rate of discount of the
Prussian Bank were all at a minimum of respectively 3^, 4, and 3j-£ per
cent. A significant rise begins in 1845, which clearly links up with the
Kondratieff processes of the forties — railroads, in particular. But it
lasts until about 1870, when yields of first-class fixed interest securities
were about 5 per cent. We might question the evidence. Requirements
of war finance and the speculative excesses of the fourth Juglar could well
be held responsible for what, from our standpoint certainly, is an irregu-
larity, and there were significant relapses. Since this irregularity is,
however, parallel to the one observed in the behavior of price level, it is
THE RATE OF INTEREST 631
reasonable to connect it with the effect on demand for balances that a
rising price level would have. The course of events in 1870 was domi-
nated by the Franco-German War, and during the next three years
interest rates fluctuated under the influence of the mania of promotions1
and the indemnity. The picture is somewhat distorted by the policy of
the Prussian Bank, which in the second half of 1872 resorted to restriction
of credit rather than to an "adapted" increase in rate. After the crisis,
both short rates and yields return to form.
English rates illustrate our mechanism to perfection. We will glance
at the market rate for high-class bills. It promptly rises at the beginning
of the forties, obviously in response to railway building, and before there
was any rise in price level. The year 1846 displays the Jugiar peak (the
annual average for 1847 is still higher — 5 pounds, 17 shillings, 6 pence2 —
but this was a panic peak) followed by the lagged decline to the second
half of 1850 (average for the year, 2 pounds, 5 shillings). Then, with the
second Jugiar rising and the Kondratieff prosperity going on, the rate
again rises — with a setback in 1852 indicative of the presence of a counter-
acting force, gold inflow — to a peak higher than that of the first Jugiar in
1856 (5 pounds, 10 shillings, the figure for 1857 is again higher, 6 pounds,
15 shillings, but still more obviously due to a panic), which marks the
culmination of the Kondratieff very well — no such annual average
occurring again within the period except in the financial panics of 1864
and 1866. Disregarding these and also the intervening year, we have a
falling tendency throughout recession and depression and only a little
upturn, as stated above, at the end of revival (from 19 shillings, 2 pence
in 1895 to 1 pound 18 shillings, 6 pence in 1897). Within this movement,
Juglars show very well — rise in 1860 and 1861; then a normal fall to
1868, interrupted by and resumed after 1863 to 1866; rise again from
1869 through a setback to the Jugiar peak of 1872 (4 pounds, 1 shilling,
9 pence, 1873 again displaying the influence of the panic in its annual
average of 4 pounds, 14 shillings); fall to 1876; recovery rise 1877, and
some panic influence in 1878; after reaction to this, a rise in 1880, 1881,
and 1882; fall to 1886; no substantial change until the lagged rise in the
prosperity of the last Jugiar (in 1889); a Jugiar peak in 1890 (average
3 pounds, 17 shillings, 7 pence); then the fall (some panic effect in 1893,
however) and the recovery mentioned above. Yield of consols tells in
this period about the same tale.
The reader will have observed that some figures have been discarded
as panic peaks, others as setbacks, a procedure which of course stands *
and falls with its historical warrant for each case. But he will also
1 We recall that from 1871 to 1873 there were 928 flotations with a capital of 2,781
million marks. The states, however, repaid debts. The Reich even bought fixed interest
securities as an investment.
* See Palgrave, op. dt.t p. 33.
632 BUSINESS CYCLES
observe that, if due attention be paid to the interference of the three
cycles with each other and to the influence of wars, panics, and monetary
disorders, little remains to be accounted for by factors other than those
that make up our cyclical mechanism, and that, in particular, the
behavior of rates within each phase of each cycle is substantially according
to expectation. This is all the more significant because of the variations
in gold production that occurred in the period. As a matter of fact, we
have seen that they did assert themselves. But neither their short-run
effect in depressing rates nor their long-time effect in raising them was
strong enough to prevail over the cyclical movements. No country was,
taking the period as a whole — Calif ornian gold, of course, acted primarily
on the United States — so directly exposed to the impact of new gold as
was England. Nowhere can actual short-run effects be so easily followed
up. Yet interest behaved in no other country so nearly according to
theory as it did in England.
3. It is — for prewar times — not possible, but it is fortunately not
necessary, to prove fully the statistical relation between interest and
profits, which, since our analysis makes the latter the dominant factor in
the variations of the former, should be particularly close under any
normal circumstances. As far as we may, however, trust the testimony
of the German dividend series (see Chart XXXVIII), it is in fact, very
close: American figures about dividends declared give the same impres-
sion; so do railroad earnings. In comparisons with other series, the
extreme emphasis that the interest rate places upon short-time surface
fluctuations in business — this is, as has been pointed out, the only differ-
ence between its cyclical behavior and that of employment; see Chart
XX — and result trends must be taken into account. The lag in interest
rate, although it does not always show significantly and although it is in
many cases difficult to identify among all those erratic elements which
make sequences in our material so unreliable, must also be borne in mind.
With these provisos, we can say that interest goes well with pig-iron
production or unfilled orders of the United States Steel Corporation,
and that, in the two longer cycles, particularly in the Kondratieff, it
displays the relation to total output which we expect on the theory : the
long periods of falling interest rate are also the periods of strongest
increase in total output. The relation is predominantly positive in the
Kitchins as they appear, if a least-square trend be eliminated from the
output series. The short-run relation of interest to building is incon-
clusive. Interest is positively correlated with outside clearings, outside
loans, and outside deposits in prosperity and depression, also — though
less so — in revival, but negatively in recession. Since the latter fact is
the main reason why there is a trend in these series, trend elimination will
again produce lagged covariation in all phases.
THE RATE OF INTEREST 633
The reader will have observed that the expectations which our model
yields for the behavior of interest are, with two exceptions, almost exactly
the same as the expectations for the behavior of price level. One of those
exceptions is the lag, which, however, is important only for short fluctua-
tions. The other is the downward trend in prices. But barring these
there should be all but exact covariation, because even external disturb-
ances and the internal irregularities — these with the exception of the
panic peaks in interest — are in most cases likely to affect price level and
interest in the same sense. This is in fact what we find, and not less
in the long than in the short run: a glance on the pulse charts will satisfy
the reader that this relation is not less in evidence in the Kondratieff
than in the Kitchin and that it persists in the most varied constellations
of external factors.1 For the period from 1886 to 1925 it has been
the subject of Mr. Zinn's most interesting study, previously quoted,2
which yields the conclusion that neither variable can be considered the
fundamental "driving force/* but that both "presumably fluctuate in
response to an unknown common set of generative causes, active at the
root of the system," and that "the interest rate at any time is related in a
systematic manner to the preceeding values of wholesale prices, as well
as to the concurrent values."
The mere fact, however, that price level and interest rate are only
two of very many — or of several aggregative — variables suffices to prove
that the deviations from strict covariation of the two must have other
than random causes. In particular, the strong inverse relation that
exists between the short fluctuations of deposits devided by price level
and the short fluctuations of rate of interest, both taken with reference
to their Kondratieff movements, suggests that there are systematic
reasons for such imperfections of that covariation as we observe,3 and
the above analysis but leads to the door of more intricate problems.
1 Of. Mr. Keynes' Treatise on Money, vol. II, Chap. 30 VIII: the Gibson Paradox. He
rightly calls it " one of the most completely established facts within the whole field of quanti-
tative economics though theoretical economists have mostly ignored it" which "ought to
be susceptible of some explanation of a general character." The reason why such explana-
tion is not to be found in the cyclical covariation of prices and interest is (p. 201) that it
is a long-period "rather than a strictly short-period phenomenon." But the long-period
covariation envisaged by Mr. Keynes is simply the covariation in the Kondratieff cycle.
From our standpoint, we had better speak of a Gibson Katadox.
2 Review of Economic Statistics, October 1927, especially, pp. 189-197. That period
has been divided into three subperiods— 1886-1899, 1899-1912, 1912-1925— accidentally
covering almost exactly different Kondratieff phases. The correlation between the factors
expressing the interest-price relation in each of them ("system factors," as Mr. Zinn sug-
gestively calls them) is as high as 0.82 for the first subperiod and the two others, and 0.90
for the second and third subperiods.
8 That idea, the result, and the method by which it was attained, entirely belong to
Dr. C. E. Thomas.
634 BUSINESS CYCLES
4. Within a system of interdependent quantities there is no point in
trying to label some as determining and others as determined. Within
a definite process, however, which runs its course in the system there is
point in asking what is the particular role of every element in the sequence
of events. In this sense we call interest consequential because it is,
within our process, moved away from its neighborhood value by the
entrepreneurial demand for balances and does not by its own behavior
cyclically disrupt that neighborhood. This proposition follows from the
analysis in Chaps. Ill and IV and cannot be disproved by a post hoc, ergo
propter hoc argument exclusively based on the fact that prosperities are
invariably preceded by moderate rates of interest. But it is necessary
to safeguard it against another line of reasoning — one that we have met
before and that is primarily associated with Professor von Hayek's
famous book Prices and Production.1 We assume perfect equilibrium
of perfect competition — coexisting, however, with a positive rate of
interest2 — to start with.
Then there will be, as our own model shows, unused facilities for the
creation of balances. Every bank has, in a perfectly competitive member
bank system, a prima-facie motive to make use of them. This will reduce
the rate of interest below its neighborhood value — the real or natural rate
in the sense of that group of authors — hence, increase the value of all
durable goods,3 the production of which will therefore, absolutely and
relatively to that of transient goods, expand in response to this stimulus
and turn out to be untenable as soon as the latter is removed, i.e., as soon
as money rate again equals the "real" rate of interest — the implication
being that this is bound to happen when the banks run up against the
technical limit of their power to create balances. This reasoning seems
to give a rational schema of cyclical movements, the motive force for
which is entirely supplied by the initiating action of banks. It will be
shown in the next chapter why it is not likely that banks, even if they
work perfectly competitively, would take that initiative which is by this
theory held to disrupt the existing state' of equilibrium. But that no
1 We are here interested in only one aspect of that theory, though a fundamental one,
and we do not intend to discuss it as a whole. For the sake of argument full employment
of resources at the outset is taken for granted. On other points see Hansen and Tout,
Annual Survey of Business Cycle Theory, Econometrica, April 1933. The argument we
are going to discuss also underlies Professor von Mises' and Mr. Hawtrey's theories of the
cycle and goes back to Wicksell.
2 This assumption is also granted for argument's sake. If the writer's theory of interest
be accepted, the problem we are about to discuss does not arise at all.
8 This formulation, somewhat more general than the usual one, is due to Professor
Machlup, who has also in a very telling way stressed the difference in the effects which
variations of interest exert on cost of current production and cost of long-time investment
(of. Interest as Cost and Capitalization Factor, American Economic Review, September
1935).
THE RATE OF INTEREST 635
such initiative can as a matter of fact be the prime mover of the cyclical
deviations from equilibrium is obvious because, though interest may only
rise with a lag, it certainly does not fall in prosperity, or at its beginning.
This would, however, be necessary in order to produce the above effect
on the system of commodity values. It cannot be replied that equilib-
rium rate must be defined with reference to the demand schedule for
balances and that, hence, if that schedule shifts upward, as it does in
prosperity right from the start, money rate may be "too low" or even,
relatively to the data of the situation, "falling," even if it actually rises.
For first, that shift would take precedence, logically as well as in time,
over any initiative of banks — which, in fact, would as an element of
explanation be rendered superfluous by it — and would have to be
explained independently, either by the action of external factors or by
innovation. And second, for old firms the demand schedule for balances
is shifted only by the effects of entrepreneurial expenditure: hence, any
expansion of their productive apparatus, i.e., expansion along old produc-
tion functions, could, independently of that expenditure, occur only if
interest fell absolutely, which it does not. It should be added that pro-
duction along the new production functions has nothing to fear from a
future rise in interest, since it is protected by the buffer of profit and
since, moreover, it so upsets previous conditions that it is by no means
certain that time preference — the real rate in the sense of these authors —
will remain what it had been. It is much more likely to fall.
But if we thus see that this theory fails to give satisfaction as a funda-
mental explanation, we also see that it correctly describes a large expanse
of fact — as has been pointed out repeatedly in the course of our historical
sketch. We do not, of course, hold that the behavior of banks has
nothing to do with the cycle. There is no doubt that without credit
creation amplitudes of cyclical fluctuations would be much smaller,
although there have been (witness England in the forties) even "manias"
with very little credit creation, and although the effect on money rates
is not the most important lever through which it works. In particular,
the phenomena of the secondary wave would then be much less in evi-
dence. Although even these are induced not by money rates being too
low but by entrepreneurial activity, higher money rates would go far
toward keeping them in bounds and low money rates tend to foster them.
Judgment in granting loans is much more important than the rate
charged, and reckless banking does not consist in financing cheaply but
in financing irresponsibly. But effects would be mitigated if, given a
certain amount of irresponsibility, credit were made more expensive.
What should be evident more than anything else is that a cheap
money policy in prosperity can have no other effect than to accentuate
excesses and subsequent breakdowns. We define it with reference to the
636 BUSINESS CYCLES
concept of Adapted Rate introduced in Sec. A. This rate, it will be
remembered, is not an equilibrating rate and does not prevent the system
from drawing further away from the preceding neighborhood, although
it does not cause its excursion. Cheap money policy, being the attempt
to keep actual rates below it, unavoidably imparts an impulse in that
direction. A dear money policy, analogously defined, would no doubt
also be effective and exert equilibrating influence. But in neither case
can effects be expected to be as great as the usual two- variable analysis
would have us to believe, for other things cannot be equal under the cir-
cumstances. Going on in the cyclical sequence we notice that the pri-
mary factors which bring about the upper turning point are independent
of the rise in rates that has previously occurred. In this sense we may
say that interest no more causes the down turn than it causes the excur-
sion of the system into prosperity. Navigation becomes, however, diffi-
cult beyond this proposition. To facilitate it we will at first disregard
the "lag" and assume that rates move synchronously with the other
relevant elements of the system. Then the rate will move pari passu
with the incipient self-deflation of business. But, even so, it will press
with different severity on different sectors of the system and certainly be
one of the factors to make the operations of the Secondary Wave, or
some of them, unprofitable. And cheap money policy would be as effec-
tive to counteract this as action on any single, though important, element
can be.1 The argument is at first sight much strengthened by the fact,
first inserted into our model in Chap. IV, that important lines of expendi-
ture, some types of dwelling-house building, in particular — but many
adaptations to the things that newly emerged in the preceding prosperity
and a great part of the conquest of the new economic space come under
the same heading — are actually highly sensitive to variations in interest
alone, which therefore here acquires, as any other element can more or
less, a "secondarily causal" role. But as our study of time series has
shown, interest actually does fall, by virtue of the working of the cyclical
mechanism, fully as much as, and more than, is necessary to keep up sys-
tem expenditure, since those lines of expenditure do, in fact, expand
in recession. The case for accentuating this effect — in practice it mostly
rests on a failure to distinguish between the processes of recession and
the processes of depression — would therefore have to rest on individual
circumstances, and also have to be strong enough to overcome the objec-
tion to keeping alive those operations of the Secondary Wave and to
preventing the liquidation of maladjustments.2
1 In formulating as above, we follow the fashion in assuming that the effect of falling
interest on the consumers' expenditure of the strata affected is negligible. This is, however,
not necessarily so.
2 Lest to some readers this should seem an unpleasant result, the writer begs to point
out that his argument can be made to imply the necessity of much more control, regulation,
THE RATE OF INTEREST 637
No such objections to cheap money arguments apply in depression
phases. But when the "crisis," if any, is over, interest does fall as a rule
very considerably and sometimes precipitately. There is no better
example by which to demonstrate how little the rate of interest can by
itself do than the typical course of events in depressions. With more,
though still little, justification could it be argued that there are junctures
at which its increase by authoritative action would be more effective in
stimulating business, because it would call forth some demand for bal-
ances, which holds back in expectation of further fall. We need not turn
to our model in order to elaborate the point. All we have to do is to
invoke common business experience which, but for the phraseology that
has developed, would be enough to establish the truth that tampering
with the rate of interest in depression is but a piece of political liturgy.
It follows that, however promptly interest falls, it would not on that
account become a major factor in bringing about the lower turning
point in business. This is independent of, but much strengthened by, the
fact that revival is, at least at first, typically a revival in current business
but not in investment, and in the cost of current transaction even zero
interest would as a rule mean much less than a moderate reduction in
wages. As revival wears on, however, interest resumes the "secondarily
causal" role it loses in depression, and cheap money policies would
become effective again. But as we may infer from theoretical consider-
ations, and as we can see statistically and historically, interest does not
and cannot rise quickly and considerably in conditions of both liquidity
and good business prospects. As far as it does rise, this rise fills an
obvious function : the effect of cheap money policy can in these conditions
only be a Hayek effect.
We have now to insert the Lag, remembering that what is usually thus
referred to is a composite of many different elements, some of which are
not genuine lags at all, since interest holds place behind other elements
in the cyclical sequence. The cases of the lag in short rates and of any
lag there is in bond yields behind short rates must be distinguished. Very
little friction, if any, enters into the first. But any failure of short rates
to rise promptly in prosperity, however caused, must tend to intensify its
primary and secondary processes — especially the latter1 — and induce a
great many operations that will lead into difficulties later. A lag in and
after the upper turning point has the effect that, with business deflating
investigation into the details of each case than any simple recommendation of cheap
money policy, hence enough of public activity and bureaucracy to gladden their hearts.
See next Chap. Sec. A.
1 It should be observed that this lag is also due to the institution of credit creation: if
prosperity were exclusively financed by savings, rate of interest would not only rise more
but also more promptly.
638 BUSINESS CYCLES
itself, a moment comes in which the adapted rate equals the equilibrating
rate and after which it acts punitively, in strict theory, for the rest of the
recession. Though this effect presupposes and does not cause recession,
it yet intensifies it. But it should be observed that we have not the
same reason to expect a lag in the turning point and in recession as we
have to expect one in prosperity. If such lag as there is asserts itself in
the shorter cycles — the case of the Kondratieff is doubtful — about equally
in all phases, this must be accounted for somewhat differently in each of
them. The main factor in recession is not risk, still less borrowing in
order to cover deficits, but the new demand which replaces that of the
entrepreneurs, partly at once and partly — later on — when the rate has
already begun to fall. It follows that that punitive effect is not of a
nature to cause general disturbances, although it does increase the diffi-
culties of the typical Old Firm. In particular, no such general punitive
effect should be inferred from a lag of short rates as against prices. For,
as we know, falling prices are compatible with flourishing business. In
depression unemployed balances ride their attack on interest rate, and any
lag in this phase can be due only to the element of risk. With decreasing
risk, the weight of those balances which must first melt before the rate can
recover also accounts for the fact that is particularly in evidence during
Kondratieff prosperities and more than anything else responsible for so
many people's unshakable belief in the efficacy of a low rate of interest,
viz., that the lower turning point of the shorter cycles so frequently occurs
when the rate of interest is still falling. Since it is business activity that
pulls it up, we shall not wonder at it. About the causal importance of
this favorable factor it is not necessary to add anything to what has been
said above.
The lag in bond yields as against short rates is not a regular phenom-
enon at all. This apart, it differs from the lag that has just been dis-
cussed, in that friction plays a considerable part in it; and it can be best
observed, on the one hand, in the upgrade of the current Kondratieff in
this country and, on the other hand, frequently in deep depression after
the panic, if any, has passed. But the first instance, like others that could
be cited, is to be explained on special grounds that have nothing to do
with the working of our mechanism. The latter instances have received
more than their due share of attention. These cases owe their existence
to the fact that in the atmosphere of such a situation, with the causation
of which interest has little to do, neither borrowers nor lenders care to
do any business at any rate whatsoever, and that the price of existing
bonds is kept down by the knowledge that the sources which service them
are being increasingly impaired. In general, short-time covariation of
yields and short rates is instantaneous.
CHAPTER XIII
The Central Market and the Stock
Exchange
A. Banks and the Pulse of Industry. — Turning to questions concern-
ing the way in which financial facilities are provided for the purposes of
providing financial facilities, accommodation for providing accommo-
dation, we move still further away from the motor forces of our process.
In doing so, we shall continue to use the general schema of the sphere of
banking which has been sketched out at an earlier stage as an instrument
that will help us to interpret the actual working of the banking systems
of our three countries. We have also at various turns of our way had
enough glimpses of how they actually function to enable us to piece
together a serviceable picture. Those traits which are most relevant to
our present purpose may be summed up and somewhat amplified as
follows.
In the period under discussion the institutional pattern of banking
systems (quality of their personnel and their traditions included) under-
went considerable changes, even since Peel's act in England, the found-
ation of the Reichsbank in Germany, and the establishment of the National
Banking System in the United States. But most — though not all — of
them were in effect, whatever the intention may have been, nothing but
adaptations to the situations created by our process and in fact part of
the latter. Such general propositions about the whole period as we are
forced to make should, hence, be reformulated for each of the historical
subdivisions, which in strictness ought to be dealt with individually.
1. The individual member bank may be so big as to be able to influ-
ence by its own action price level and money rates, either by the mechan-
ical effect of its operations or by its example. The member bank business
of the Bank of England is an outstanding case for about 1845-1870. In
the last three or four decades of our period the great English and German
concerns, toward the end of it also the New York group which was dubbed
Money Trust, afford other instances. Whether a bank holds such a
position or not, however, it finds itself hedged in, not only by familiar
conditions which, though elastic in the long run, yet technically limit
639
640 BUSINESS CYCLES
the sum total of balances it can put at the disposal of its customers at
any given point of time,1 but also by the fact that it cannot normally
take the initiative in lending or even, without taking the initiative, allow
loans mechanically to expand to that limit. This simple fact is so much
covered by the embers of ancient and recent controversy and so liable
to being misunderstood that we must explain our meaning with some
care, even at the risk of repetition. We will first speak of member
banks' business with their customers, then of their operations in the open
market, and finally, of a particular aspect of the investment item.
When nearly a century ago, in the controversy on the principles of
banking that centered around Peel's legislation, Fullarton held the view
which was put forth again and again, before as well as after, by writers
representing the views of the banking world — namely, that banks cannot
"force their money upon people" — he laid himself open to the rather
obvious reply that they have a strong motive (in perfect competition, at
least) to use their facilities to the full and that, hence, they will so frame
their conditions as to call forth the corresponding demand. This reply
fails to do justice to what Fullarton really meant and but repeats the very
error that he probably wished to point out, viz., the error implied in the
application to banks of the general schema of business behavior. Any
other manufacturer simply wishes to sell the quantity of product which
will yield the maximum net revenue and does not bother about what
happens to the units after he has sold them. The manufacturer of bal-
ances who wants "his money" back, cannot for this and other reasons
behave according to the same schema. For him other considerations
enter into every transaction with every customer and make of it an
individual case, which cannot be dealt with in the same way as the sale of
a pair of boots. Moreover, every one of these individual cases is, on the
one hand, an element of his relation to the customer which must be
viewed as a whole, and on the other hand, an element in his total position
which, also, must be watched as a whole. This forces upon him an atti-
tude of resejrve, which is entirely absent from the behavior of any other
businessman. To be sure, this attitude is not always observed. We
have met examples in our historical report. The fact that often it is
not observed is the very reason why there was also plenty of practical
wisdom in the teaching of the so-called currency school (Lord Overstone)
which Fullarton fought, and which in turn fought his teaching, according
to the approved method of economists, which consists in never meeting
1 This distinction between the limits set at any time to the expansion of the deposits
of all the member banks taken together and to the expansion of the deposits of any single
bank is part of traditional teaching and, hence, need not be elaborated. It should, how-
ever, be recalled that traditional doctrine stresses inadequately, first, the elasticity of
those limits, and second, the fact that since individual banks expand and contract together,
the limits set to the expansion of an individual bank lose much of their importance.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 641
one's opponent's point. Nevertheless, that attitude is an essential
part of the logic of a banker's situation. And it follows that, while most
shopkeepers will normally congratulate themselves whenever they are
"sold out," the banker does not typically aim at being, and does not
congratulate himself if he is, "loaned up." On the contrary, this means
for him an exceptional and undesirable situation of embarrassment and
of danger, which is, as a matter of practice, always recognized to mean
that, both by the individual banker himself and by the banking com-
munity. Customers' business cannot be handled safely and comfortably
unles^ each bank has a generous allowance of unutilized lending power.
Full utilization of that lending power in member banks* business with
their industrial and commercial customers is, hence, no equilibrium con-
dition— nor is it an "adapted" condition outside of neighborhoods of
equilibrium — for the banking system, and it cannot be said to be an
obvious interest of the banker, supplying the explanatory principle of
his behavior. Economists who, on the strength of a general schema of
business behavior, insist that it is have themselves to blame for the quotes
that the practical banker applies to their "theories." This consideration
completes, from the practical side, an argument presented in the preced-
ing chapter, and, of course, applies with added force to the only real
case — that of imperfect competition.
Now, initiative may mean many different things. It is not suggested
that a bank's directing committee is an automaton. Our view of banking
is, in fact, much less mechanistic than that of those theorists who attri-
bute to banks the role of prime movers in business cycles. This is
obvious from the emphasis placed throughout our analysis on the element
of purpose. For us it is neither the formal character of the business to
be transacted — e.g., the discounting of commodity bills of exchange —
nor the security that makes sound banking, but knowledge and under-
standing of, and proper attention to, the purpose which the balances
applied for are to serve. Judging the chances of success of each purpose
and, as a means to this end, the kind of man the borrower is, watching
him as he proceeds and granting or withholding further support accord-
ingly— these are the fundamental functions of that committee which are
more important than the mere decision how far the bank is to go in
granting loans, how much, if at all, it should lean on the bankers* bank,
how great a risk of maneuvering itself into a tight corner it should under-
take, and so on. The above statement that a bank cannot normally take
the initiative in its business with its customers merely meant that it can-
not normally initiate the individual transaction. Its truth becomes
obvious if we consider what that would involve. As far as the financing
of enterprise is concerned, it would involve suggesting definite plans to,
or urging on, people who have every motive to go ahead and must nor-
642 BUSINESS CYCLES
mally be expected to know the ground on which they are standing better
than any banker can. There are, no doubt, exceptions to this. The
practice of banks of the credit mobilier type supplies the most important
one, especially in cases like mergers or in cases in which an innovation
acquires additional importance from simultaneous developments not
within the horizon or radius of influence of the man who is to carry it out.
Independently of this there are occasions — they may be the high points in
the career of a great banker — in which a bank can successfully make itself
responsible for an enterprise by pledging its support and committing
itself to seeing the entrepreneur through. But it is evident how risky
this is. As a rule, it augurs ill for a proposition if it has to be forced on
the primarily responsible man. But the same applies, in a lesser degree,
even to current transactions. A bank often sees reason to restrain, but it
is rarely in a position to ask its customer: Won't you borrow in order to
do this or that ? The chief exception with some banks in some countries
is the role they assume in dealing with private investors or speculators.
But that is another matter.
It might be urged that initiative need not go as far as this but can be
exerted without any particular suggestions by assuming a general attitude
of encouragement, which consists mainly, though not wholly, in offering
attractive conditions and in conveying to wavering customers the impres-
sion that if they go ahead they will not do so alone. Even a small bank
may do that if it observes that others do it, but a bank important enough
to influence situations by its own action or example seems particularly
able to impart such a stimulus. Hence the question, which in fact is very
frequently asked, why the banking system does not use this power or even,
as some would put it, why it uses it viciously, i.e., so as to intensify both
booms and depressions. Recalling our discussions of historical cases
and the relevant argument in the preceding chapter, we may first repeat
that no encouragement from banks is necessary in order to start a pros-
perity phase and that such encouragement as has been actually proffered
— it mostly comes within our concept of reckless banking, witness, for
example, the events preceding 1837 and 1907 — can be shown to bear a
close relation to the occurrence of crises and downward spirals; second,
that no discouragement from banks is necessary to bring about the turn-
ing point into recession in our sense, and that such discouraging influence
as is sometimes exerted only serves at that juncture to steady and not to
dislocate the system; and third, that near or at the lower turning point
interest rate is in any case low and falling and bankers' attitude as a rule
encouraging enough. That question, therefore, narrows down to the
behavior of banks before and during depression. It has been pointed out
in the preceding chapter that the end of recession is the only stage to
offer possibilities of corporative initiative by banks. The reason why
THE CENTRAL MARKET AND THE STOCK EXCHANGE 648
more use is not made of these possibilities is analogous to the reason why
initiative action by banks could do but little once a downward cumula-
tive process has set in. Banks control one element only of the situation
in which businessmen find themselves, and that situation contains so
many hitches and untenable positions that such action, besides becom-
ing extremely difficult, cannot be expected to be effective. But it is
true that the survival interest of each individual bank then drives it into
courses of action which, even without panicky calling of loans — very
important though that element is in the mechanism of crises, and though
it was especially so during the earlier part of the prewar period — tend to
intensify the phenomena of the spiral. All this only amounts to recog-
nizing again, this time from the side of banking practice, that interest is
not the hero of the cyclical drama.
No apologetic purpose is either intended or served by this analysis.
On the contrary, its suggestions could, if this were within our purpose, be
worked up into a very comprehensive plan of regulation; but it would have
to be primarily restrictive in nature, although not wholly so. And it
would have to aim at improving personnel and at enforcing adherence to
standard practice rather than different principles of practice. Machinery
which would enable some authority to force banks to take initiative action
would, if effective at all, in most cases lead to additional maladjustments.
In those cases, moreover, in which such machinery could be expected to
have remedial or preventive effects it would still be inferior to a policy
that, while leaving the banks free to fill their function, would directly
act on the economic process — by government expenditure, for example.
2. To economists who can see in the business cycle nothing but an
effect of the working of the banking system, and particularly in depressive
situations nothing but "deflation," which could almost at will be turned
into prosperity by " refl ationary " creation of deposits defeating the
attempt of business to deflate itself, our argument may well be inaccept-
able, although it seems to agree with rather obvious facts of business
experience — which is the reason why we fared so well, in discussing the
behavior of clearings, deposits, and interest, with the working hypothesis
of a "passive" banking system.1 Nor will those agree for whom pros-
perity and depression differ in nothing but the state of businessmen's
minds — elated in the one condition and depressed in the other. For if
there were nothing in the objective situations to account for those humors,
it would be plausible to assume that psychotherapeutic behavior on the
part of banks would never fail to be effective. However strong the
1 It should be observed, however, that the word passive does not express our meaning
well and might easily prove misleading. The preceding paragraphs sufficiently show that
we do not mean more than that banks primarily act in response to business situations,
which cannot in turn be explained by their action.
644 BUSINESS CYCLES
evidence which verifies our view, it cannot be made fully convincing
except by reference to the whole of our analysis of the cyclical process.
But some points still remain that require specific notice. They all link
up with member banks' operations in the open market.
It has been stated above that while the behavior of banks during or in
expectation of breakdowns may acquire causal significance for some of the
features of "crises," the upper turning point is not normally brought
about by banks' calling loans because, having run up "against technical
limits, they are losing cash and, therefore, have to retrace their steps.
"The greater number of cyclical fluctuations keep well within these
limits," even in this country, and there is "no basis for the belief that
these cyclical swings never halted until the resources of the banks had
been exhausted."1 But banking systems seem to approach those limits
much more closely than they actually do, because they do not keep the
whole of their surplus funds in the form of cash and balances with other
banks but allow part of them to earn interest by means of temporary
investment. In substantial agreement with practice, we call such
investment the Secondary Reserve. Now this policy is, indeed, what
the customers' business is not, viz., first, actuated exclusively by the
desire to use funds as much as possible; second, entirely a matter of the
initiative of the banks' managing committees; and third, as mechanical or
routine as anything in business can be. As soon as committees have
settled what credits they are to grant to their customers, the surplus to be
temporarily invested, in all normal situations and excepting cases of
speculative booms that make the open market abnormally attractive, is
determined and will be invested — as far as member banks are concerned —
with little regard to the general business situation, which asserts itself
more in the kind of investment to choose than with reference to the
decision to invest.
This is, as we have seen in Chap. XII, what pulls down open-market
rates so drastically in depression as soon as panics, if any, are over, and
even in the other phases normally tends to keep them at a low level. As
we have also seen, effects extend to the stock exchange and in particular
to bond prices. But there they stop until business reacts to more funda-
mental stimuli. Even bond issues do not react to low open-market rates
alone. Conversely, progressive liquidation of temporary investments
when business revives, restores normal financial habits and helps to
raise rates again. However, even if banks were always "loaned up" in
this sense, the same consequences would not follow as if they were loaned
up in the relevant sense, i.e., in their customers' business. In particular,
they need not normally, if "loaned up" in the wider sense, restrict the
latter on the ground that they have expanded to a technical limit —
1 A. A. Young, op. cit., p. 28.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 645
although as prosperity wears on other reasons may suggest themselves for
doing so — and the very presence of secondary reserves outside the depres-
sion phase is proof that they never want to be "loaned up" in the narrow
sense.
What secondary reserves consist in is a question of fact to be answered
differently for different times and countries. Where there are facilities
for rediscounting, high-grade bills of short usance, short -time government
paper, and so on are the classical items; where such facilities are lacking,
call loans to the stock exchange. It depends on the organization of
markets and on the traditional behavior of the public what other "embod-
iments of future balances" qualify for the role. Government bonds,
possibly also other bonds, may so qualify; and, where banks have the legal
power of holding it and trading in it, even common stock. In the case
of German banks, the motive of employing surplus funds in this way
undistinguishably merges with the motive of supporting the prices of
securities of the concerns they patronize. It should be observed that
each member bank thus fulfills with reference to firms and households and
within its sphere of influence a function akin to that which bankers'
banks' open-market operations fulfill with reference to member banks.
It buys, and thereby supplies the public with liquid funds, on the down-
grade; it sells, and thereby curtails the public's available means — a qualifi-
cation will presently be noticed — on the upgrade. This mechanism does
not work alone or in a world in which other things are equal and therefore
its regulative effects are, no doubt, often overshadowed by many other
factors. But the profit motive is sufficient to set it into motion.
In the case of common stock, this is obvious. In the case of bonds, the
practice under discussion means that banks buy when bonds are dear and
sell when their prices have fallen — accentuating the fluctuations in bond
prices which would otherwise prevail. Such operations are remunerative,
nevertheless, since the funds engaged in them would otherwise be idle.
So we see, again, that neither machinery nor motive is absent for meeting
any pessimistic moods of the public by the creation of balances. Only,
there is little point in arguing that it should be made more effective, since,
as pointed out in Chap. XI, even the balances actually created tend to
become as idle as the corresponding funds of the banks would be.
There can be little doubt that the investment item of the American
banks was for our period mainly of the nature of secondary reserve.
There was a significant special trend in it. But both New York and
outside banks — which do not differ as much in their investments as they
do in loans — purchased bonds primarily when their surplus funds
increased and sold — or, at all events, purchased less — when their cus-
tomers' business expanded. In the case of outside banks, though not of
those in New York, this meant strong inverse relation in the fluctuations
646 BUSINESS CYCLES
in investments to the fluctuations in loans and discounts, which suffices
to establish our point.1
3. Of course, member banks' investments cannot entirely be accounted
for by the secondary -reserve hypothesis. To begin with, political neces-
sity or pressure may, in times of abnormal government expenditure,
compel banks to increase their investment in government bonds to an
extent altogether beyond the range of secondary-reserve considerations.
Second, we know that one method of financing real investment, or even
current expansion, is to sell assets to banks (Chap. XI), an operation
which differs from borrowing in technique only, but tends to impart to
the investment item a time shape exactly opposite to the one we have
described. The statistical picture proves that for the prewar period the
latter prevailed. This need not always be so, however. During and
after a time of abnormally great government expenditure, for instance,
not only banks but the public may be so saturated with government
bonds that buying and selling them may well become the central element
of their dealings with each other. In this case, firms and households
would, on the one hand, be much more independent of normal bank
credit and, in consequence, of the advice and the approval of their banks,
and it would be much easier for them, or some of them, to resort directly
to the open market in order to finance themselves. On the other hand,
the banks, while losing their hold on industrial operations, would be
much more in a position to act initiatively within the sphere left to them
under such circumstances. They could then with more justice be said to
"regulate the flow of funds" and to control the volume of balances.
The postwar period illustrates this.2 Only it would not, on that account,
be any truer to say that they thereby regulate the pulse of business.
On the contrary, the momentous change which permanence of such
conditions would imply — and developments since 1930 strongly suggest
that these conditions, and in particular government expenditure that is
abnormal from the standpoint of capitalist logic, may have become
permanent by now — precisely consists in paralyzing such regulative
influence as they had beyond mere regulation of the amount of balances
which may or may not be used. An essential piece of capitalist mecha-
nisms would be gone forever. It is easy to see that it would be
1 The behavior of notes in circulation accords perfectly with our view. Barring the
influence of the introduction of the 2 per cent consols by the Act of 1900, they fluctuate
much as money outside of treasury and banks and with the rate of interest. They thus
followed the pulse of customers' business, as we should expect.
2 In spite of the sentence that follows in the text, it should be pointed out that, since
most adherents of what we have called the Investment Theory of Banking (Chap. Ill)
as a matter of fact mainly reason on postwar facts and problems, the difference between
our view of banking and those of most economists of our day is not quite so great as it
seems.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 647
government, acting not through the mere offer of facilities but through
expenditure, and not the banks, which might then really acquire something
of the power that is being wrongly, though usually, attributed to the banks :
these would be more powerless than ever in starting the movement of the
economic body from troughs to higher levels of activity, though they
would have the same power they always had to foster excesses.
Third, the investment item transcends secondary-reserve considera-
tions and acquires additional importance also in the case of banks which
on principle participate in the ventures they finance — roughly, in the
case of the crSdit mobilier type,1 as exemplified by German banks. It is
part of the task of any bank's managing committee to exert what, in
distinction from business initiative, we may term financial initiative,
i.e., to suggest to, sometimes to impose on, its industrial customers
financial policies and in particular answers to the questions how far and
how long they are to finance themselves by borrowing in current account
and when and to what extent they are to fund these debts by the issue of
bonds or stock.
As we have seen (Chap. XI), such refunding either annihilates
deposits or, if the bank (or another bank) finances subscriptions, sub-
stitutes new debtors for the old one, and in both cases powerfully affects
not only the situation of the bank itself but also the whole banking system
and the open market. Issues intended to finance real investment may
be said to do the same thing potentially. But banks of the mobilier
type themselves acquire parcels of stock, either because this strengthens
their hold on the current business of customers and their claim to manage
the future issues of the same concerns, or because they wish to influence
the latter's business policies — particularly to acquire positions of strategic
value in negotiating amalgamations — or because a given issue is not yet
ripe for introduction to the public, or because they intend to trade in it
permanently, or simply in order to profit by appreciation. Such invest-
ments, which in many cases develop into the most valuable part of
the assets of a bank and into the backbone of its business, and which enter
the balance sheet either as "permanent participations" or as "securities"
(Effekteri), are, of course, built up for their own sake and are largely,
though not wholly, exempt from the considerations incident to what
English or American opinion holds to be normal banking business.
The statistical implications of this need no emphasis. Other aspects have
been discussed in Chap. VII. As has been pointed out there, even in this
case the real influence of banks does not, in general, amount to control
1 It should be borne in mind that the above expression is used because the writer sup-
poses that it conveys what he means. We know that historically it is not quite correct.
See Chap. VII.
648 BUSINESS CYCLES
of what is being done, although it often amounts to control of the share-
holders' meeting.
B. The Central Market (in an Isolated Domain). — First, the limited
purpose of the following discussion of the role of banker's banks in the
cyclical process must be borne in mind throughout. A very misleading
impression would be gathered by the reader if from the points to be touched
he tried to construct a general theory of central banking. Second,
complete analysis would have to consist of a string of historical inter-
pretations of individual central banks or central bank systems, some frag-
ments of which have been offered in previous chapters. Beyond a small
number of propositions that may be true of every institution or class of
institutions that can be partly or wholly identified with central banking,
it is extremely dangerous to generalize, because environments, organiza-
tions, habits of doing business, attitudes of and toward member banks,
firms, and speculation, relations to government, and so on differ so widely
as to impart different meaning, at different times and in different coun-
tries, even to identical forms and phraseologies. Two instances will be
sufficient to emphasize this once more. A universal spirit of sound
money and sound banking in the classical sense pervaded the English
business and banking communities for most of our period. This made all
the difference for the nature and effects of any measures taken by the
Bank of England: exactly the same measures — and, of course, both
policies and phraseologies of English banking were widely copied — taken
by a bankers' bank in a country of different mentality would have
differed from the paradigma, both in economic meaning and in effects.
Again, we shall later notice the importance for the policy of the Bank of
England and for the results it produced, of England's creditor position.
For the moment we defer consideration of gold movements, international
relations, and exchanges, and consider central banks as if they functioned
in isolated domains. But even so, the difference of which that creditor
position was one of the consequences, remains essential, viz., presence at
any given point of time throughout the period of a relatively great amount
of accumulated wealth which behaved according to an established
tradition and made things very much easier for the central institution
than they were in other countries. This fact alone suffices to make the
application of the principles of English banking to other countries as
doubtful a matter as the application of English principles of parlia-
mentary government to other national patterns has been or would have
been.
Third, our central market, consisting of the transactions between
the bankers' bank and member banks and the operations of the former
in the open market, is, of course, a bold abstraction. From previous
discussions it should be obvious that in the case of the Bank of England,
for instance, our concept of transactions between it and member banks is
THE CENTRAL MARKET AND THE STOCK EXCHANGE 649
(in part) a useful fiction representing actual fact. But it is (in part) a
fiction all the same. Moreover, until the foundation of the Federal
Reserve System, hardly any central bank confined itself to bankers'
bank business m our sense, and no central bank — not excepting the
Federal Reserve System — ever covered all the functions we attribute to
bankers' banks. As regards the latter point, we ought really to speak,
even in the English case, of a bankers' bank system rather than of a single
Central Bank. As regards the former point, the historical importance
of central banks' member bank business has been emphasized before.
It should be added that this business was a most important lever for
influencing the structure of credit and greatly strengthened the hands
of the central bank in its bankers' bank functions: it gave it a hold on the
market which any mere bankers* bank must always find it difficult to
acquire. That in some respects it also fettered the hand it made stronger
will, on reflection, be seen not to involve any contradiction.
Finally, in the American case, some readers have probably taken
offense before this at our associating central -bank functions in this country
primarily with New York banks. This seems to run counter to obvious
facts, the outstanding one being that under the rule of the National
Banking System as amended in 1864 and 1887 a system of reserve and
central reserve city banks developed, which legislation itself, by imposing
special reserve requirements — in the case of central reserve city banks,
also, by insisting on reserves' being kept entirely in their vaults — recog-
nized as bankers' banks and balances with which other banks were allowed
to count as part of their legal reserves to % of the total amount of the
latter. This practice of using checking accounts with certain banks as
reserves, or of redepositing reserves and, generally, surplus funds with
correspondents in more central positions, of course, established a relation
of current cooperation between the latter and their banking customers
(brokers, bond houses, investment trusts, and so on as usually included).
Correspondents undertook to act as agents for these as for other cus-
tomers, advised them, did their foreign-exchange and open-market business
for them, used them in turn as agents for collection and other pur-
poses, and, last but not least, accommodated them in cases of seasonal and
cyclical tension. This bankers' bank business was highly competitive, and,
hence, worked out very advantageously for the member banks in our
sense of this term. They received, not only a considerable amount of gra-
tis service, but also interest on some rough average of their credit balance,
which for the two last decades before the Federal Reserve Act is usually
estimated at 2 per cent.1 According to the Comptroller's reports, about
1 On the other hand, the correspondents, unless they confined their business with non-
bank customers to local proportions, were hardly able to go without these connections.
This fact, the not very profitable character of these connections in themselves, and the
650 BUSINESS CYCLES
half of the deposits of the banks in central reserve cities were bankers'
balances. Under these circumstances, our reliance on New York bank
figures, of course, means relying on a sample of American bankers' bank
activities. But this sample includes the roof of the structure and its most
important elements, and as has been mentioned before, the behavior of the
figures goes far to justify our choice.
We start from the analogy that we have previously noticed between
the relation of a member bank to its customers, firms and households,
and the relation between a bankers' bank and the member banks it banks
for. Once the nature of the variations in system expenditure in the
cyclical process of evolution and their relation to the balance-manu-
facturing activity of member banks (member creation, as we call it) is
fully understood, the fundamental questions of principle regarding the
balance-manufacturing activities of bankers' banks (central creation)
are, ipso facto, solved. In particular, we readily realize that a bankers'
bank can, still less than a member bank, afford, or have an interest, to be
"loaned up." This and the consequent fact that central banks always
kept a comfortable margin are for England and Germany so obvious that
it would be waste of space to elaborate them. In the United States, both
facts are blurred by their very recognition by legislation.1 The tech-
nique of this legislation, however, so curiously mixed up quantity and
safety considerations that it would have been necessary to keep a surplus
reserve, even if the legal minimum had been sufficient to meet all cases —
which a rigid figure in the nature of things can never be as long as it is
below 100 per cent. But this was, owing to the circumstances just
glanced at, both a difficult and a costly thing to do. However, it is worth
while noting that, quite abnormal situations excepted, the New York
banks' resources were never exhausted in any phase of the cycle, and that
they almost always managed to keep free surplus reserves, though only
modest ones — in the last decade before the war they averaged for the
months of the autumnal drain at about 1.5 per cent.
This, of course, quite accords with our view of the actual role of cen-
tral banks and indirectly teaches something about it. Their position
necessity to supervise closely the member banks on the safety and success of which the
safety and success of those bankers' banks themselves largely depended, combined with
the doubtful quality of the managements of many of the — especially pigmy — member
banks to create a situation, the natural remedy for the practical difficulties of which would
have been mergers plus branch banking. The merger movement, in fact, gathered momen-
tum from 1900, when the amendment to the National Bank Act still intensified the struggle
between larger banks for member banks' balances (see chart on p. 358, H. P. Willis and
J. M. Chapman, The Banking Situation, 1984). If it did not go far enough to strengthen
the credit system of this country, that was due to hostile legislation and to an attitude of
the public mind which would not have stood the formation of units big enough to be really
safe and effective.
1 Statutory requirements as to reserves against bank notes have nothing to do with the
above argument.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 651
and the quantitative weight of their business are such as to make their
mere attitude and example a matter of moment to the banking com-
munity and to the business world in general. Rediscount rate — or, more
generally, the conditions at which they stand ready to finance member
banks — rationing, suasion, open -market operations undertaken ad hoc,
need be mentioned only, to call up roughly appropriate ideas in any mind
that is not misled by theories and plans. It may, however, be added that
rationing was not an emergency measure but part of the ordinary routine
of bankers' banks. No member bank was allowed to rediscount or
borrow all it may have wished to, but, if the writer may judge from the
practice of the only great central bank which he thoroughly knew and
from such indications as he was able to gather about the practice of two
others, every member bank was closely watched, not only as to its balance
sheet, but also as to its personnel, the nature of the transactions it entered
into, its affiliations, and the kind and quality of its customers — whether,
for instance, they were retailers or industrial firms, geographically and
economically distributed or concentrated, and so on. Gossip about them
and their leading men was carefully collected. Thereupon, it was
decided what its "ration" was to be. This was then varied cyclically,
besides being currently revised on the merits of the individual case. By
suasion, as distinct from (general) attitude, we mean attempts to influ-
ence individual member banks or groups of member banks. It covers a
wide variety of things and would, in some cases, better be designated as
scowling or snarling. It ranges from threats — of which the threat to
withhold accommodation is only one — and admonitions down to such
measures as extending or withholding invitations to official dinners.
Though the possibility of producing effects by such methods greatly
differs from country to country and from time to time, no one who is at
all familiar with the working of credit institutions will deny that it is
considerable. This alone suffices to prove, if proof be wanted, that
bankers' banks cannot harbor any systematic tendency toward being
•"loaned up," and that the argument in the previous section applies to
them with added force.
It would be quite wrong to think that central bank action is merely
adaptive or passive in any sense other than that it is typically action in
response to a situation not itself created by the central bank. The writer
confesses his inability to see the point at issue in the ancient controversy
whether bank rate is "declaratory" or "constitutive." As far as the
bankers' bank also transacts member bank business, it cannot help being
if not the, at least a, leader in an imperfect market. As far as it does
central bank business only, it is in certain cases a monopolist and in all
cases in a position in which any of its moves must be a factor in the sub-
sequent situation. Even if it never aimed at anything else but at fixing
its rate in such a way as merely to declare or register what the financial
652 BUSINESS CYCLES
situation is, we should have to recognize that the man who has it in his
power to hoist a danger signal at will thereby unavoidably acquires the
power to create, as well as to declare, a situation. But the very position
in which a bankers' bank finds itself makes it practically impossible
merely to register the state of things such as it is at the moment, still less
the tendency of market rates alone.
Central banks' rates and attitudes owed much of their effectiveness
to the fact that member banks and the public looked upon them as
symptoms of the situation and, to a considerable extent, reacted accord-
ingly. It is also true that the powers of central banks to exert mechanical
effect seemed — to economists as well as to business men — to be greater
than they really were. Yet the means at the command of central banks
to exert such effects, sometimes acting upon sensitive margins, were
obviously adequate to managing any ordinary situation and also to
securing for them considerable freedom of action, though much more of
course in the case of single central banks than in the case of competitive
bankers' bank systems. Criticism, which has become traditional, draws
a different picture — mainly, it seems, for three reasons.
First, most critics concentrate attention on exceptional situations.
It is, no doubt, relevant to the analysis of the functioning of any central
bank or central banking system, to see how it lived in rough weather.
But to make this the only criterion implies overlooking all the problems
to be solved in the normal run of things, which it is easy to overlook
precisely because they were solved successfully. Second, it has been
urged that central banks reacted mechanically to the behavior of certain
indices relevant to their own situation and perhaps that of the money
market and of the currency, without any regard to the economic organism
of their nations as a whole and without any conception of the therapeutic
influence they might have exerted. But if we consider what it was that
the Bank of England was supposed to react to — reserve proportion, gold
movements, and so on — it becomes evident that in the vast majority of
cases (and in exceptional situations neither the Bank of England nor the
Reichsbank acted upon such indices alone), diagnosis of the cyclical
situation of a country would have coincided with the inference that any
outside observer could have drawn from those indices, and that action
upon that inference must, hence, have had some "stabilizing" effect on
volume of transactions, prices, and so on. Finally, however, most
critics speak from the background of a banking theory of the business
cycle and imply that central banks should be able to iron it out entirely.
Such a misconception of course negatives all possibility of realistic
analysis and substitutes a spurious problem for the real one.
It is at the same time obvious from the nature of the means by which
central banks may influence economic processes that there are definite,
THE CENTRAL MARKET AND THE STOCK EXCHANGE 653
though elastic, limits to their effectiveness, and that those means go
much further if the intention is to put on brakes than if the intention is to
stimulate. That this was so as a matter of historical fact cannot well be
doubted. But it is, again, a distinct question whether this was so merely
by virtue of the ideas about sound currency that framed the legal powers
and duties of central banks, their traditions, and practices, or because of
more fundamental reasons. As we have seen in the case of member
banks, that legal or customary restrictions, though they may be prac-
tically unavoidable to curb recklessness, do not create the necessity for
that attitude of restraint that is inherent in the conditions of banking,
so we have no difficulty in recognizing in the case of central banks, that
an analogous attitude would impose itself on them by virtue of the
logic of their situation, even if no legislation or tradition enforced it.
The main obstacle that prevents the adherents of the most varied
"schools" to see this is the argument that a central bank, being the
"ultimate creator of credit," is exempt from the limitations banking
practice imposes on member banks, and hence, in the absence of those
legal restrictions and traditions, would enjoy almost unlimited freedom
in acting on business situations. Quite apart from the fact that govern-
ment could never afford to allow it to fail, it need not bother about
"quality" and "purpose" of credit at all, since it would have it in its
power to improve any "quality" and to justify — in the business sense —
any "purpose " to any desired extent by further creation of balances. We
will discuss this with reference to the sequence of cyclical situations in an
isolated domain, although our material does not permit the elimination
of the complications that arise from international trade, gold movements,
foreign exchanges. At every step, problems of actual and of possible
behavior will be kept distinct. The factor of redcemability of balances
in a legal tender that also serves as cash enters, of course, into the former
only.
As pointed out above, the central bank may, owing to its conspicuous
position, exert some influence on the temper of the business community,
which at certain critical junctures may be considerable — especially in
preventing or stopping panics — but ordinarily cannot go further than
hortatory efforts by leading officers of state, which are immediately put
into their place by the recognition of the truth that cyclical phases are
more than mere matters of psychology. Excepting this and the influence
exerted by a central bank's member bank business with firms and house-
holds, it is clear from the outset, however, that its action in the central
market affects the money and the open markets primarily through mem-
ber banks reserves,1 and hence the business processes beyond those mar-
1 It is not required that the regulation of member bank reserves be a definite, clearly
perceived, and theoretically understood aim of the managers of the central bank. Such
654 BUSINESS CYCLES
kets only insofar as variation in these reserves — supplemented by atti-
tude and suasion — can influence them. But this means, first, that, even
if central banks were so completely masters of member bank systems that
their slightest move would be immediately translated into an increase or
decrease of members' willingness to lend, no greater effect on industry
and trade could be expected than we have seen reason to expect from
variations in this willingness: one only of the data on which firms act
being affected thereby, definite results could be predicted only under
ceteris paribus assumptions, perfectly inadmissible within our process.
Second, even for member banks themselves increased and decreased
power to lend is only one of many factors that determine their willingness
to lend, and the decisive one in some situations only and not in all.
Member banks are distinct centers of economic decision. No policy of
the central bank, short of a declaration to sanction and make good the
consequences of any action whatsoever, can alter all the data on which
such decisions rest and, in particular, the logic of the banking business
described in the preceding section. It would, hence, not even be correct
to say that the central bank determines the "supply schedules" — if this
expression were permissible — of member banks, although it influences
them by its greater or lesser readiness to shoulder the burden of tensions
that may arise. The analogy between the relation of a member bank
to its industrial customers and the relation of a bankers' bank to mem-
ber banks is, therefore, not, as one might think, destroyed if the bankers'
bank is the ultimate creator of credit in the sense in which a single central
bank may be said to be: inferences from the associations which that
phrase is likely to call forth would be completely misleading. Nor do
open-market operations alter the argument. It has been often observed,
and is indeed a patent fact, that member banks occasionally thwart the
action of the central bank — a qualification will, however, be added to this
in Chap. XIV, Sec. F, V — and that the latter sometimes proves to be a
"dud." This is perfectly understandable from our standpoint and is no
deviation from, but part of, the blueprint of the credit machine. The
fact that central bank action sometimes does and sometimes does not take
effect is a problem — and an awkward one — only for believers in the
mechanical action on business of mere volume of credit facilities. What
central bank policy can be said to determine directly — pari passu with
members' reserves — is the central-market rate. It also very powerfully
acts on open-market rates. But there is a long way from action in the
central market to action on the money market in our sense, and a still
regulation will in many cases be implied in their ordinary business behavior. The Bank
of England exerted regulating influence of this kind on country and on London banks, at
least from the beginning of the nineteenth century, without developing any theory or
principle of it for a long time. Open-market operations are about as old.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 655
longer one from this to action on business activity and prices. Even
legally unlimited power to create balances does not imply actual power to
create them, still less the power to make them active.
Since central banks mainly act on business through member banks,
it is unnecessary to show that the former, no more than the latter, "cause"
the system's prosperity excursions, though they could conceivably prevent
them. In strict theory, particularly if the preceding neighborhoods come
sufficiently near to being equilibrium states, central bank action at such
times may with equal justification be said to be absent or neutral or con-
fined to those small adjustments of the steering wheel that are necessary
even on the straightest course in the calmest of seas. In the absence of
effects of external factors or of phases of underlying cycles, there is, during
this situation, as much encouragement to member banks in the central
bank's attitude and rate as there is encouragement to entrepreneurs in
member banks' attitudes and rates, and no more — there is hence no
Wicksellian effect from any discrepancy between money and real rate of
interest. The presence of external factors — capital and gold movements,
in particular — may alter all this, and it is in connection with these that
the true role of central bank initiative reveals itself. But this we dis-
regard for the moment. As prosperity sets in, the central bank typically
watches the expansion of business volume and of banking activity,
slowly raising its rate in response. Normally there is at this stage no
motive to exert restrictive initiative, nor would it be easy to do so, both
because of the pressure of public opinion and because of the fact that the
money market is, as we have seen, not only sloppy in the sense in which
the market of any commodity may be, but in the special sense that its
funds are normally underutilized in and immediately after a neighborhood
of equilibrium, so that any practically feasible attempt at restriction
would be thwarted by members,1 and nothing but the most drastic open-
market operations could avail — borrowing unlimited amounts at 10 per
cent, for instance.
As prosperity wears on, however, and member banks liquidate part of
their secondary reserve — calling demand loans to the stock exchange and
to bill brokers — the rising bank rate tends to become effective, both in
the sense that it applies to a larger volume of business which the tighten-
ing of the open market drives towards the central bank, and in the sense
that it exerts restraining or punitive influence. In London, for instance,
more and more of the bill brokers' material had, at that stage, to be
offered at the Bank for rediscounting. This theoretically goes on, and in
1 In practice member banks could also thwart a policy of stringency which the central
bank may be trying to impose, by shifting balances among themselves and drawing funds
from abroad. Issues of stocks and bonds, however, primarily act in the direction in which
the central bank wishes to steer. A policy of restriction may, in fact, be intended to enforce
656 BUSINESS CYCLES
practice mostly did go on, until the situation reversed itself and the
economic process settled into the perfectly normal and anything but
catastrophic processes of recession. There was no spectacular manage-
ment about this, such as ardent stabilizers wish to see, but it was mone-
tary management involving a lot of general guidance of things, and
central banks, such as the Bank of England and the Reichsbank, can no
more be said to have mechanically followed events than a horseman, who
refrains from spectacular whippings and spurrings and acts on his mount
primarily by small adjustments of his seat, can on that account be said to
"follow" his horse. Nor were there at those junctures any sudden jerks
caused by central banks' running up against technical limits and causal to
the occurrence of the upper turning point — which could be compared, if
we wish to keep to our analogy, to brutal pulls at the curb, productive
of discomfiture to horse and rider.
This is what, in the absence of external disturbances, central banks
actually did, although the picture would in its details have to differ for
different periods and countries, in particular according to whether central
banks directly rediscounted for member banks or not. It should be
observed, first, that in important points, though not in all, our sketch
supports the classic theory of central banking. But it could almost
equally well be expressed in terms of more up-to-date theories of monetary
management. The objection to both is that they remain on the surface.
But as far as they go, the real difference between them does not justify
the emphasis put upon it. Second, it should be noticed that central
bank behavior as outlined also conforms to the rules the profit motive
would dictate. Only, they would have to be defined not by the principle
of maximization of instantaneous profits but with reference, on the one
hand, to the general logic of banking and, on the other hand, to the fact
that a concern, situated as the central bank is, can never fail to link its
interest with the state of the whole economic organism. Third, the
formal analogy may be mentioned which exists between the cyclical
and the seasonal situations that central banks have to face. When, in
particular, prosperity is under way they face a situation technically
similar to that which first attracted the attention of economists under the
heading of Autumnal Drain.1 This and similar seasonal variations in
the volume of borrowing and in the flow of cash (income tax payments,
Christmas business, and so on) bankers' banks learned how to handle,
although the fact that crises broke more often at times of seasonal strain
than at other times persisted throughout the period. In England there
was no significant association between bank rate and either London
clearinghouse figures or circulation (notes plus coin) within each year,
1 Cf. W. S. Jevons, On the Frequent Autumnal Pressure in the Money Market, Journal
of the Royal Statistical Society, 1866.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 657
although there was between seasonal gold movements into and out of the
Bank of England, bank rate and the price of gold. In Germany the
handling of seasonal strains was facilitated by special legislation (increase
of tax-free amount of bank notes issued in order to cope with quarterly
requirements; no provision, however, was made for the autumnal drain).
In the United States the experience of 1907 exerted some pedagogical
influence that relieved seasonal tensions. But more important than the
analogy is the difference between seasonal and cyclical problems — the
former presenting nothing but the clear-cut task of minimizing incon-
venience; the latter, the much more formidable one of dealing with a
fundamental economic process.
What central banks could do if they were freed from technical fetters
and, in particular, from the pressure of currency movements and of
foreign exchanges is another matter. It has been stated above that they
could, in the limiting case, not only prevent all member creation but,
by taking up all available funds, prevent prosperities themselves. It is
no less clear that in the atmosphere of high prosperity, with excesses
everywhere in full swing, expanding member bank reserves would induce
increased lending and take effect in the business world. Hence, by
unreservedly standing for indefinite inflation bankers' banks no doubt
could defer occurrence of the upper turning point indefinitely, i.e., until
the monetary system breaks down. Unless this be an end in itself,
there seems, however, no point in trying to fight recession when it sets in
if we remember what it normally means. We have here the final reason
why any really or hypothetically unlimited power of central banks to
create credit makes, practically and analytically, less difference than one
would suppose.
When, in the course of recession, business normalizes itself and liquid-
ity increases all round, central banks' control over member banks'
reserves and the open market and their indirect influence on business
become progressively weaker. The horse is no longer running up to his
bit. But in the absence of international complications this does not
normally create any problem necessitating interference by restrictive
open-market operations. On the contrary, central banks are in this
phase both able and willing to give rein and even to urge on by a low
rediscount rate, although no initiative on their part is as a rule called for.
This statement, which implies that depression (not crises) should set in
while conditions of easy money prevail, hence not because of monetary
stringency, can sound paradoxical only to adherents of monetary theories
of the cycle — some of whom will, it is hoped, be converted by the obvious
facts of 1937 to 1938 — and amounts to saying that central banks can do
little beyond what their admonitions may effect in order to prevent
depressions. What can be done at all from the banking angle, member
658 BUSINESS CYCLES
banks should in that situation be able to do themselves. This is almost,
but not quite, independent of the legal fetters imposed on central creation,
for if there were absolutely no limits to the latter and if the central bank
made itself unconditionally responsible for whatever transactions people
were prepared to enter into and undertook to finance all ensuing deficits
ever after — otherwise a crisis would immediately follow upon the with-
drawal of the guarantee — this would no doubt produce effects.
For the same reason and with the same qualification, there is not
much room for central bank initiative in the course of depression. If
facilities were forced upon them by open-market buying, member banks
would thwart the intention, first, by using their access of funds in order to
repay their debts, and then by accumulating secondary or simply idle
reserves. Moreover, it follows from our diagnosis of the nature of
depression that, as far as they could be prevailed upon to accept against
their judgment such business as they may be able to attract under the
circumstances, this would be a source of additional difficulties in the
future. But there is another way in which the latter effect may follow
from an attempt of bankers' banks to enforce expansion of the
volume of balances during depression. It is obvious how difficult,
if not "politically impossible," it must be to retrace such steps. The
opportune moment for doing so will, in the eyes of the member banks and
of the public, never come. The delicate processes of incipient revival
could, in fact, be easily stalled by anything looking like restriction, and
public resistance is not likely to decrease later, particularly if member
banks have accumulated not idle but secondary reserves. Thus the
system might reach the neighborhood of equilibrium in a state of abnor-
mal liquidity, and in the subsequent prosperity the central banks' action
in the previous depression, or the member banks' surplus reserves created
thereby, will indeed produce effects, viz., speculative excesses, reckless
banking, an overgrown Secondary Wave, and later on, breakdowns. It is
this vicious effectiveness of therapeutic efforts of this type rather than
the mere futility of trying during depression to drown pessimism in a
flood of credit that should be stressed in arguments about it. We may
note in passing that permanent expansion of the circulating medium
may in this way ensue, and that fluctuations of prices may thenceforth
be permanently from a level higher than would otherwise obtain.
Gold discoveries act in the same way rather than by directly creating
prosperities.
It is recovery which, of all the phases of an ordinary cycle, presents the
most difficult and most important practical problems for bankers' banks.
The situation may so clear up in the last stages of depression that the
responsibility can be taken for a lead that may help to bring about the
lower turning point more quickly without producing undesired effects
THE CENTRAL MARKET AND THE STOCK EXCHANGE 659
as well as the desired one. But for reasons we need not repeat, such a
lead is likely to be followed much more promptly by those sectors of the
community which are likely to overdo things than by the ordinary run of
"legitimate" business. And this difficulty grows in importance in the
later stages of recovery, when everyone realizes the upward tendency
and is disposed to press forward in any case. The necessity for steadying
advance thus soon becomes more evident than that for propelling it,
and we shall not be surprised to find that central banks were, in revivals,
as a rule more concerned with controlling general liquidity than with
trying to add to it, and to take an attitude which the public very naturally
disapproved. Under the general conditions of revival, the money and
open markets were most likely to get out of hand, and it was, hence,
primarily the experience with the excesses and vicissitudes that fre-
quently occurred in revivals — Juglar revivals, in particular — which
eventually led to an almost general demand for measures calculated to
strengthen central bank control. It follows that central banks could have
done more than they did to stabilize things in revivals. But the main
obstacle to this was public opinion, although until the end of the seventies
neither the task nor the technique was perfectly understood. They also
could, if freed from technical fetters and from any other consideration
except how to produce booms, have accelerated the processes of revival
at the expense of producing corresponding slumps.
Nothing has been said about how central banks did behave or, to
retain our double-track argument to the end, could have behaved in
crises. In order to bring out the fundamental question, the above analy-
sis has been conducted within the walls of assumptions that not only
excluded the problems incident to international relations — gold move-
ments in particular — and to domestic disturbances of a non-cyclical
nature but, as much as possible also, the abnormal features of the cyclical
process itself. Within these restrictions we have seen that, by guiding
and managing all the time, central banks allowed the cyclical process to
take its course, taxing expansion in the later stages of prosperity, but
did not by their action create any of the cyclical phases; that such guid-
ance had also another rationale than mere regard to golden fetters would
imply; that it rarely if ever required that bank rate should precede in the
time sequence of the factors of business situations and that its failure to
do so did not cause or aggravate slumps; that by its nature and not only
by statute or tradition, regulation primarily meant restrictive regulation;
but that this regulation was never mechanical or uniquely determined by
obedience to a few indices. And it would be easy to understand on these
considerations alone that bank rate was, in both England and Germany,
regularly above the rates of the open market. But it might be urged
with some justice that the behavior of such a regulating agency cannot be
660 BUSINESS CYCLES
described in terms that exclude precisely those irregularities which accord-
ing to some economists primarily call for regulation. l Within an isolated
domain, the relation of central banks to what we call reckless banking,
speculative excesses, fraudulent or irresponsible business activity —
especially, finance — is the one of the two most important points, and the
treatment of crises or panics is the other.
Dealing with the former would, in order to be effective, require a
policing power, which to this day has always been quite> beyond central
banks. This inability of capitalism to police itself is as striking as its
inability to protect itself — it always requires both a policeman and a
protector of nonbourgeois complexion, who regulate, shield, and exploit
it. This is as true of the times of Queen Elizabeth as it is today. But
it is largely this inability that produces crises as distinguished from mere
depressions. For the prewar period, at least, historical evidence is
conclusive on this point. As soon as we accept this evidence we had
better stop talking about the causation of crises in terms of bank rate or
inadequate accommodation imposing restrictions that were relaxed only
"when the damage had been done" and that turned into catastrophes
what, with another central bank policy, would have been perfectly
normal situations. Crises would have occurred if no central or member
bank had ever called a single loan, and must be understood in the light
of the fact that in any economic and social system which is unable to
prevent irresponsibility and misconduct correction by consequences is
the only method to prevent indefinite aberration. At least in historical
retrospect and on the understanding that from this recognition of patent
facts alone no practical conclusions can be drawn as to possibilities and
desirabilities of later times, this should be admitted, however much we
may understand, and sympathize with, all the strata which that method
victimized. Then both the initiative more or less regularly taken by
central banks in such situations and its limitations will appear in their
proper light. Their position was such that the social meaning of crises
translated itself for them into obvious business considerations. Owing
to their inadequate power and will to police the world of banking and
finance, their action came, indeed, unavoidably after the event — this is
the only sense in which it can be held that it came "too late," and this
sense has nothing to do with bank rate. Coming when it did it had to be
remedial and not punitive. Success went in some cases, which we met
in our historical sketch, to the length of preventing panics and that
1 Although the present writer does not go so far as that, but admits that even for the
most normal course of things there is a special case for monetary as distinct from other
regulation, it should be pointed out that the difference between the views presented above
and some that are at present more widely held is partly due to the fact that we deal here
with one particular aspect. Many views on monetary management, for instance, that
seem to differ fundamentally from ours primarily refer to the international gold situation.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 661
restriction of credit which will ensue in panics. In most other cases, it is
easy to see that consequences were substantially mitigated and durations
of spirals shortened, as compared with what we may conceive they would
otherwise have been, by the initiative action taken by central banks.
The general logic of their position shows in such situations, as it were
through a magnifying glass. They obviously could not simply stand for
saving everything and everybody, because this would have meant con-
doning error and misconduct to the point of seriously impairing the
efficiency of the system with which they had to work. Saving the
Gurneys, for example, would have meant saving ironworks and shipping
ventures which could not have lived without permanent subsidies. But
within the limits set by this consideration the helping hand was freely
extended. In the English case in particular, the suspensions of Peel's
Act must be interpreted not, as contemporaneous and later criticism
often interpreted them, viz., as so many breakdowns of the engine, but as
parts of it that were put into operation in order to show in moments of
panic the "ultimate creator of credit" in all the glory of unlimited power.
That engine was not foolproof, to be sure, and the way in which it was
operated at any given point of time and in which it was made to deal
with situations that were essentially historical individuals may on a
different level of analysis give plenty of scope for criticism, quite inde-
pendent of any theories about the magic possibilities of bank rate. But
this does not affect the question of principle, with which we are concerned.
Thus, while we need not stay to discuss the obvious arguments that
may be adduced against, or the equally obvious case that may be made
out for, that institutional arrangement, or the question how its sources
of waste compare with those of the Gosplan, it is necessary to emphasize
that within its framework central banks could hardly have done much more
than, for example, the Bank of England or the Rcichsbank actually did.
Control over business beyond those limits cannot be usefully discussed in
terms of banking policy but only in terms of a much more thoroughgoing
type of management of the underlying industrial and commercial proc-
esses, which would call for agencies of different construction. In this
sense bank reform is a technical and, moreover, an intracapitalist matter,
which cannot be dealt with in an extracapitalist spirit and which has
little to do with the fundamental issues of today.
Chart XXXV is presented in order to illustrate the results of this
analysis. The behavior of Notes in the Hands of the Public, of course,
reflects, besides the cyclical process, the secular or "structural" change
that occurred in their role in the monetary system. A similar remark
applies to Private Deposits, which otherwise, together with the ratio of
Banking Department Reserve to Deposits and Bank Post Bills and with
Other Securities, conform to expectation from our argument to an extent
662
BUSINESS CYCLES
which is remarkable, considering how many external factors there were to
produce deviations. If the reader looks at the chart in the light of our
historical report, he will easily satisfy himself that our process substan-
1850 I860 1870 1880 1890 1900 1910
CHART XXXV.— Bank of England and Allied Series (see Appendix, p. 1063).
tially suffices to account for the behavior of those series. London Total
Clearings and the ratio between them and Total Deposits (of joint stock
and private banks as reported by the Economist) facilitate interpretation
since the middle nineties. The ratio between London Bankers' Balances
and Reserve obviously contains a special trend.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 663
How far our instance of a competitive system of bankers' banks, the
national banks of New York City, qualifies for the role may in part be
inferred from Chart XXXVI. It should be added that the growth, as
evidenced by the increase in capital and surplus, of New York national
1875 1880 1885 1890 1895 1900 1905
CHART XXXVL— United States (see Appendix, p. 1064).
1910
1914
banks after the reform of 1900 — much stronger than the increase in
capital and surplus of outside national banks — was precisely due to their
success in attracting deposits of national and state banks outside New
York and, hence, to bankers* bank business. Study of the chart reveals
the fact, pointed out before, that both net deposits and loans of New York
banks moved in the short run inversely to the deposits and loans of out-
664 BUSINESS CYCLES
side banks.1 This implies inverse association, in the short run, of New
York loans and commercial paper rate. As has also been mentioned
already, there is less difference in behavior in the case of investments,
though the strongly inverse short-run association between outside loans
and outside investments is not present in the New York figures, and in
the case of reserve money held. The distribution of money, " lawful " and
other, between New York and outside banks does not seem to have
played any considerable role in the mechanism of short cycles in this
country. All this substantially holds also for Juglar fluctuations, but
in the largest contours the differences are naturally much less in evidence.
We conclude that New York banks in fact acted as bankers' banks.
Of course it must be borne in mind that bankers' bank business was only
part of their total activities. But it is nevertheless clear that the varia-
tions in their loans and investments were a function of the flow of money
— both of cash and of deposits — from and to their "member banks,"
which cyclically swelled and depleted their deposits. Interpretation is,
hence, exactly opposite to what it was in the case of outside banks, in
the one case deposits, in the other case loans being the primary phe-
nomenon.2 Net deposits reflecting the variations in the difference
between the amounts due to and due from other banks were, in New
York, persistently and considerably larger than loans and fluctuated more
strongly than in outside banks.
The competitive character of this bankers' bank system, of course,
deprived its component units of much, if not all, the freedom that a single
central bank enjoys. Much less power for initiative action was left to
them, and they were often compelled to sail very close to the wind. As in
every fall deposits were withdrawn and cash streamed out into the West
and South, so in every prosperity an analogous phenomenon asserted
itself and forced them to liquidate their temporary investments and even
to take gold from Europe against bills drawn in payment for exports or by
special arrangement. After having emerged from the Kondratieff
depression in a liquid position, they had a difficult time from the second
half of 1886 until the last quarter of 1893 and the first half of 1894. Then
they recovered elbowroom, but in their situation they could do little to
keep it. Nor were they able to fortify their situation systematically and
to plan ahead, as the Reichsbank did after 1902. Hence, after five years
of easy steering, their surplus cash vanished in the course of the events
1 The comparison with the graphs for outside banks is laborious. Also, short-run
fluctuations are much better revealed if trends are eliminated. For both reasons, the
reader is invited to inspect A. A. Young's charts, op. cit., pp. 26 and 29.
2 Since inflow of money immediately created deposits and a corresponding access of
cash but not of loans, the difference between deposits and loans must show in the surplus
reserve item. So it does; see A. Piatt Andrew, Financial Diagrams, No. 22.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 665
we have glanced at in Chap. VII. Responsibility for what no doubt
was a poor showing was discussed there. It can not simply be attributed
to the institutional arrangement. In particular, the argument that this
arrangement threw the burden of cyclical tensions on the New York
banks and purchased comparatively smooth working of the outside
banks at the price of unstabilizing the former seems to miss the point.
For this is precisely what bankers* banks are for. If it be held that they
were unequal to the task, it should be emphasized that what they were
unequal to was not the handling of normal cyclical situations but the
prevention of those abnormal excesses the causes of which were too
deeply rooted in the psycho-sociological pattern of the time to be con-
trolled by any central bank, but the spirit of which had infected them also.
The competitive character of this system of bankers' banks was, how-
ever, not its only peculiarity. Perhaps still more important was another.
Owing to the absence of an effective rediscount mechanism and other
specifically American conditions, New York banks were in particularly
close relation to stock exchange speculation. Everywhere member banks
lend to the stock exchange, and everywhere bankers' banks thus indirectly
helped to finance both new issues and speculation. New York banks
would have done so even if they had been pure member banks. But in
addition they applied to this purpose by far the greatest part of the funds
that came to them from their member correspondents, so that financing
the New York stock exchange was the direct complement of their bankers'
bank activity, and stock exchange call loans stood in the place which in
Europe was occupied by surplus cash or by secondary reserve items of a
very different character. Often even time loans were made to serve the
purpose. If member banks lent surplus funds directly and used their
New York correspondents merely as agents — which was, as a rule, the
more profitable method — the situation, while substantially the same,
admitted of still less initiative on the part of the bankers' banks. No
comment is necessary;1 but the reader should allow himself to be impressed
by the close, though not perfect,2 covariation of New York Clearings,
both with Value of Stock Exchange Transactions and with Stock Prices,
which our chart displays.
1 Nor is it necessary to comment on the practices concerning the "float" and "gross
deposits" which it would be so enlightening to discuss for the purposes of a fuller diagnosis
of the nature of that situation.
2 Not only do imperfections occur as to trends and on many individual occasions, but
there is also in the raw figures of clearings a seasonal, which is absent in stock prices. After
all, New York industry and commerce was bound to assert itself. The Kondratieff
branches show equally well in all series — the "breaks in trend" — though with unequal
emphasis. So they do in loans and deposits. Juglar phases differ. Value of Transac-
tions rises from 76, Clearings from 78, for instance. Call rate, on the whole, agrees better
with Value of Transactions than with Clearings.
666 BUSINESS CYCLES
C. The Cyclical Aspects of International Relations. — These aspects
of international relations cannot receive due attention within this book,
and very few remarks can be offered in addition to what has been said on
various occasions in the historical chapters. Of all the limitations
imposed by the plan and purpose of this book, this is the most serious
one. Not only do cycles in different countries systematically affect each
other, so much so that the history of hardly any one of them can be
written without reference to simultaneous cyclical phases in other coun-
tries, but cycles really are, especially as regards the great innovations that
produce the Kondratieffs, international phenomena. That is to say,
such a process as the railroadization or the electrification of the world
transcends the boundaries of individual countries in such a way as to be
more truly described as one world-wide process than as the sum of dis-
tinct national ones. Capitalism itself is, both in the economic and in
the sociological sense, essentially one process, with the whole earth as its
stage. Both reasons — interactions and supernational unity of funda-
mental processes — explain why in our historical survey the cycles in our
three countries were found to be so much in step.1 The fact that they
were is not more obvious than the mechanism that produced it and also —
in principle, at least — the manner in which these relations affected the
working of prewar central banks and of the prewar gold standard.
1. Even if international economic relations consisted of nothing but
trade in commodities and services, the cyclical behavior of exports and
imports, as has been pointed out before, could not be expected to be as
regular as that of other aggregates. Not even as to trend can any general
proposition be formulated, for the process of capitalist evolution may
work, and at some times and in some countries actually has worked, in
the direction of increasing, instead of in the direction of decreasing, the
autarky of nations, quite independently of any policy aiming at the former
end. In a cyclical movement fashioned according to our pure model
expectation would, if that movement were confined to one country and
if the economic process in the others were stationary or merely growing
(in our sense), be for decrease of exports and increase of imports in the
positive phase, and for the opposite behavior of both in the negative
phase. We cannot hope to find this, of course. But traces of it show
in many instances, so for this country from 1872 to 1878, in 1881 to 1882,
and in 1907. More convincing than totals, however, are imports and
exports of non-agricultural commodities. Hence, business situations in
the other countries would in this case tend to display an inverse "cycle."
If the innovations which are responsible for a given cycle in a given
country directly act on foreign countries, opening them up, for instance,
1 On this, see W. C. Mitchell, Business Cycles, the Problem and the Setting, Chap. IV,
Sec. V.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 667
as markets in which to buy or to sell, other expectations would, of course,
follow. If the cyclical process is general, in the sense that all the coun-
tries that trade with each other display a cyclical movement of their
own, relations become much more complex. In the simplest case in
which the cycles are synchronous and all innovations national, i.e.,
such that they do not directly affect, by rivalry or complementarity,
foreign industrial and commercial structures, effects on exports and
imports become a question of relative intensity of phases. But if direct
interference with foreign industrial processes superimposes itself on these
effects, we get a rich tableau of possible cases, development of which
seems to the writer to be one of the most important of the reforms of
which the theory of international trade stands in need.1 The reader
should experiment with this suggestion.
We will confine ourselves to observing that although, in general, a
behavior of import and export series must be expected to result from
this which at first glance may seem erratic, yet conditions and relations
of any given country are sufficiently stable for sufficiently long periods to
make it possible for us to get along in many cases with quite simple
theoretical patterns. This stability shows, for example, in Professor
Taussig's famous study on British terms of trade by the almost perfect
inverse association of the variations in the Board of Trade's wage index
with the variations in net barter terms (equal to ratio of price index of
imports to price index of exports).2
From the standpoint of a central bank, that part of international
commodity movements which comes about in response to changes in
cyclical phases is in some respects a corrective and in other respects a
disturbance of the process with which it has to deal. Any increase in
imports in prosperity and any increase in exports in depression would be,
and actually often has been, a stabilizing — respectively, restraining and
supporting — influence. But, in general, the impact of foreign, and the
repercussion of the effects on foreign countries of domestic, innovations,
prosperities, and depressions will incessantly affect terms of trade, quan-
tities produced and in the course of being produced, employment, credit
requirements, and, if we assume unfettered gold standard all round, cash
items and reserves, in a way which, whenever international relations are
important enough, currently disturbs the domestic situation much as
noncyclical or noneconomic disturbances would. The point to be made
stands out still more clearly if we assume, first, that the domestic banking
system also contributes to financing the foreign parties to the trade or
1 This has been recognized and to a considerable extent accomplished by Mr. Spencer
Pollard in an unpublished manuscript.
2 The change in Great Britain's foreign trade terms after 1900: Economic Journal,
March 1925, see chart on p. 6. For this country, see C. J. Bullock, J. H. Williams, and A.
S. Tucker in the Review of Economic Statistics, July 1919.
668 BUSINESS CYCLES
else a foreign financial center contributes to financing the domestic trade,
and, second, that the trade in every individual commodity or group of
commodities acquires an inertia of its own, which in the short run makes
it all but independent of what happens to the trade in other commodities
or groups of commodities.1 Three things may ensue in consequence: the
domestic business situation may acquire a complexion completely at
variance with the phase of the domestic cycle — foreign war demand may,
for example, turn a depression into a violent boom; the state of credit
may be at variance with the actual business situation prevailing; and
the central bank may be unable to act in the way in which its diagnosis
of the domestic situation would otherwise induce it to act — it may find
itself dependent on conditions in some foreign country. It is at such
junctures — which mere interaction of national cycles might produce —
that central bank action is most nearly "initiative" with reference to the
domestic situation.
2. But this analysis is inadequate — and so, in this respect, is the
general theory of international trade — because it bases international
relations on commodity trade, which in turn harks back to primitive
barter and to which financial transactions are in principle ancillary.
We need not go into the question how great the sector of reality is or was
for which this model could be considered to be satisfactory. For the
great mass of transactions that make up the commerce of nations in the
capitalist epoch it is clear, however, not only that the stage has at any
given time been set for them by the cyclical process of evolution, but also
that priority in the mechanism of economic relations between nations
belongs not to trading but to finance. It would be truer to say that
modern commodity trade followed and complemented capital trans-
actions than that the latter arose out of and complemented commodity
trade. Selling presupposes lending or "capital export" in other forms,
and commerce develops within environments first created and incessantly
reshaped by entrepreneurial and capitalist ventures.
For our purpose it is sufficient to consider the effects of capital move-
ments on the cyclical situations and policies of banks, especially central
banks. We distinguish long-term and short-term transactions and again
transactions which arise from the business sphere — hence mainly from our
process — and transactions which do not — mainly public borrowings, which
1 This is the element of truth in the statement that commodity trade, even if the only
form of international economic intercourse, need not balance, because it depends on the
wants of the trading countries for one another's products. No defense of this statement,
which in most cases is indicative of a failure to grasp the elements of the theory of inter-
national trade, is intended here. But if we make the assumptions that are necessary in
order to enable us to accept the classical description of an equilibrating mechanism acting
on incomes and prices through gold movements, we readily perceive the possibility that
its working may produce an otherwise unmotivated slump or boom.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 669
we assume to be independent of the cyclical process, although this is no
more than an expository simplification — and which hence impinge on
money and open markets at random. Suppose that a banking house,
say, in the London of the late nineties, floated a bond issue which was to
finance, for example, electrical enterprise in Argentina, that subscriptions
were partly financed by bank loans, and that — in order to relieve our
argument from a few obvious propositions — the proceeds were intended to
be spent in the borrowing country. The borrower acquired a balance
in London and could, if he had wished to, have acquired the balance in
Argentina — which was, under our assumptions, what he really wanted —
by converting the English balance into notes of the Bank of England and
these into sovereigns which he might have shipped and deposited with his
Argentine Bank. Further assuming that the transaction was big
enough, his spending would have imparted an impulse to the business
situation in Argentina which, in turn, would have supplied the reason why,
on its access of cash, the Argentine bank would have been not only willing
but also able to expand its loans, and so on. But in England such a
transaction, or a sufficiently important bundle of such transactions,
might have brought down the credit structure by suddenly enforcing
violent contraction. It is true that foreign borrowers could not have
hoped for attractive conditions, or issuing houses for success, in a situa-
tion that was already tight for other reasons. It is also true that member
banks, since they were normally not loaned up, and the Bank of England,
which always kept surplus funds, would, in general, have been able to
mitigate the shock. In reality, moreover, our borrower would not as a
rule have wished to spend all the proceeds of the loan at home. Even
if he had, he would not have drawn gold as assumed, but would gradually
have sold exchange on London. Equilibrating effects on the preexisting
commodity trade would eventually have worked themselves out through
changes in incomes or in incomes and prices. The preexisting commodity
trade itself and its financial complement always contained sloppy nooks
and crannies. Assets which the lending country owned in the borrowing
or other countries could have been drawn upon to some extent. Thus,
the process worked on lines all of which were studded with additional
shock absorbers.
But the facts will remain, first, that during a period of indefinite
length a net disturbing influence on the lending country would almost
unavoidably have been exerted by such transfers on capital account,
since there was no equilibrating mechanism that would have been fully
effective, except in the long run ; and, second, that the disturbance was in
this case due to the international gold standard — which, contrary to
fact, we assumed also for Argentina — for without it Argentina might
have done just as well with a corresponding volume of domestic "infla-
670 BUSINESS CYCLES
tion" as she did with the gold. Hence, the gold standard, whatever its
merits or demerits may be in other respects, here created a problem by
throwing on the central banks of gold countries a burden that did not
arise from but interfered with their own economic process. No great
importance is claimed for that type of international transactions; but an
element of what the example was intended to stress entered into almost
any, and the consequent situations enforced much more initiative action
by the central banks than any that could conceivably^have arisen from
pure commodity trade. We may now add issues of foreign governments1
— which in the time of economic liberalism were not always dependent
on the consent of either the central bank or the foreign office of the lending
country, and which sometimes created balances that put the debtor
temporarily in a position of considerable power — and repercussions of
foreign disturbances of both economic and extraeconomic nature. It
then becomes still clearer that a capitalist country's financial and business
situation was at any time the result of a national (cyclical) and of an
international component, which were largely independent of each other
as to causation and, hence, likely to combine in an erratic way.
The very logic of a central bank's position and its own interest made
it imperative to try to coordinate the two, i.e., to protect the domestic
process from being upset by international transactions and disturbances
without impairing the foreign business of the nation, and in particular to
manage, if possible in advance, those states of liquidity and stringency in
the market which under the circumstances were not infrequently out
of step with the phase of the domestic cycle. It is but commonplace
to say that the endeavor to do this — under the conditions set by the
gold standard — became the chief motive of central bank policy, in
England understandably more than anywhere else, and that the struggle
of central banks for control over the money market (and for the freedom
of action such control would entail) was also chiefly motivated by it.
This commonplace is, however, not without diagnostic value, in view
of the emphasis put by so many economists on the role that central banks
do (or ought to) play in the cycle per se. This is, as we have seen, not
the crux of the matter. But owing to the presence of that other com-
ponent, monetary management of an entirely different type and of much
larger scope imposed itself nevertheless — in the heyday of laisserfaire —
and would have imposed itself, though to a lesser extent, even without
the strain put on the working of the gold standard by what we have called
(Chap. VII, Sec. E) neomercantilist policies. Moreover, it was this
1 In markets smaller than the London market the current operations of foreign govern-
ments arising both from their long-term financing and from the trade in commodities were
sometimes a source of major trouble. The operations of the Russian government in Berlin
during the tenure of office of Wyshnegradsky and Witte afford instances.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 671
component that made gold movements so important to central banks.
The inflows and outflows attributable to the cyclical component would,
in themselves, have been a very minor consideration. But gold move-
ments induced by international capital transactions were not only symp-
toms of a lack of coordination of the two components but possibly
dangerous in themselves. They so regularly called for action that it was
easy to draw the conclusion that all the central banks minded was their
reserve.
But the art of central banking precisely consisted in minding the
nation's business and in so steering between the possibilities of catastrophe
in domestic affairs and of discomfiture in international affairs that the
compromise would be tolerable to both. Considering that, owing to
England's huge capital exports and international interests in general,
the smallest mistake could at times have produced disastrous conse-
quences, actual performance, working with so small a gold reserve, was
truly remarkable. The resulting irregularities in the behavior of foreign
exchanges and gold movements, neither of which displays consistent
cyclical patterns, cannot be here discussed. A few remarks on various
procedures chosen by the Bank of England will, however, suffice to give
a rough idea of the extent to which we must be prepared to find deviations
from expectation based upon our model in monetary and banking series.
It is clear from the outset that these deviations cannot, except tem-
porarily, have been very important, because otherwise our interpretation
of those series by means of that model alone could not have accounted
for general contours as it did. This in itself testifies to the success of the
Bank. For it implies that the Bank succeeded in managing external dis-
turbance while retaining freedom to act with respect to domestic situa-
tions according to the principles discussed in the preceding section.
In the first place, it has to be recognized that what foreign financing
primarily imposed was increased caution and the necessity of preventing
liquidity from engendering tightness. This, together with the fact that
in a system, which works its funds so scientifically, open-market rates
will, whenever the situation of the moment permits it, immediately tend
to fall to a minimum, explains why the action of the Bank was almost
always in the direction of steadying or raising open-market rates and why
controlling the open market came to mean tightening it. This did not
necessarily mean that it wished to make money dearer for domestic
business. On the contrary, measures were sometimes taken to avoid that
when action on the latter was not cyclically indicated, as, for instance,
during the crisis of 1907 and in the notable case of 1910. l Tightening the
1 The simplest way to acquire such grasp of the matter as can be acquired from reading,
is to peruse the volumes of the Economist and the Statist. But the comments offered in
both, especially the former, however sound from the standpoint of a world that was resolved
to play the capitalist game, should not be uncritically accepted.
672 BUSINESS CYCLES
open market without tightening the money market (in our sense), or
adjusting liquidity to the ruling cyclical phase, may, in fact, serve as a
formula by which to express one type of those moves by which the Bank
of England — and similarly also the Reichsbank — attempted to coordinate
the two components distinguished above. Undesired pressure on domes-
tic business was no doubt exerted all the same, but to an extent only
which in the light of the analysis of Chaps. XI and XII and of the
preceding section cannot be held to have produced any major effects.
In the second place, it is easy to see why, as far as central bank policy
aimed at coordination in this sense, measures as to bank rate1 were
distinctly secondary to open-market operations and to a peculiar kind of
suasion. The Bank borrowed in the market — or from the London banks
directly so that, their loans to the market being curtailed, the latter was
driven into the Bank2 — or sold spot and repurchased on account, or
sold — much more rarely bought — outright, primarily in response to con-
ditions that arose from the international business, in order to be able to
deal with the domestic situation on its merits, although, of course, in
many cases both classes of considerations pointed in the same direction.
The peculiar kind of suasion consisted in securing the cooperation of
individual holders of big balances, such as India House or, for a time, the
Japanese government. In the paradigma above discussed one of the first
things which it would have occurred to the Bank to do would have been
to convey to the Argentinian concern the impression that the Bank had
plenty of means to make things in the future more or less comfortable
for it, and that it was just as well not to withdraw gold in an inconvenient
way or to create, by lending in the open market, a state of liquidity the
Bank did not wish to see. The gratitude of the Bank was an asset that
it may have paid to acquire at the price of some sacrifice of interest.
In the third place, if the Bank merely desired to strengthen its own
position without interfering with domestic trade at all, the device of
varying the purchasing price of gold was resorted to with considerable
success. Finally, there was the possibility of special arrangements with
foreign central banks, particularly the Bank of France. A well-authenti-
cated instance is afforded by the two transactions, the one with the
Russian Bank (i.e., the Russian government), the other with the Bank of
1 Those measures, in turn, consisted in manipulating not so much the official rate as the
rate on advances or in rediscounting at a rate higher than the official one and in similar
devices. All of them amounted to discrimination with respect to purpose, which on a few
occasions at least was also resorted to directly. Cases of the last-mentioned type illustrate
our argument particularly well. For it was foreign paper — American, especially — or paper
serving foreign purposes, that was discriminated against.
2 This, for example, was done in December 1905, when it was intended to put on brakes
also in the domestic business; see the Economist's annual survey of 1905, where it is, how-
ever, stated — incorrectly, as the writer thinks — that this was a "new departure."
THE CENTRAL MARKET AND THE STOCK EXCHANGE 673
France, during the Baring crisis. Others are more doubtful.1 For us
only two things matter: never was such a measure actually taken or
"suspected" except in connection with the international component of
English situations, and in each case the motive was to protect the working
of the domestic business organism from being disturbed by foreign dis-
turbance or by repercussions of foreign disturbance on English finance,
which is the reason why our series were not more affected; and there is no
justification, either for the patriotic irritation displayed by the Economist
and some English writers at the suggestion that that measure was taken
several times, or for the opinion of other writers that it each time spelled
a breakdown of the prewar bar) king policy. In the internationalized
world of the free gold standard, borrowing by one central bank from
another was a perfectly natural thing to do and nothing to be ashamed of.
Management of the international component, however, was so success-
ful only because there was the powerful wall of short and semiliquid
claims on foreign debtors that sheltered the English structure. Without
it, that structure could never have been worked on such small margins of
safety or by means of such delicate adjustments. This mass of claims —
Mr. Hartley Withers estimated it in 1909 at from 150 to 200 million
pounds — which was currently turned into cash to be presently reinvested
almost anywhere within the gold area, responded to the Bank's slightest
move very much more promptly than foreign-owned balances would have
done, facilitated great capital transactions, supported foreign business,
mitigated domestic stringencies. Because of its presence, tightening the
open market — raising open-market rates — not only regulated but, by
drawing gold, eased situations. Similarly, cyclical increase in bank rate
had not only restrictive but also relieving effects,2 and incidentally — it is
quite misleading to make this the center of the matter — turned unfavor-
able exchanges, as a rule, with the utmost ease. Hence, we cannot be
surprised to find that a general belief in the effectiveness of bank rate
grew up among bankers and economists which, without this short capital
situation, it would be difficult to understand. It is more astonishing that
1 Explicit arrangements, however, were not always necessary. It is easy to see that,
in some cases in which "assistance" might have been thought desirable, it would at the
same time have been to the interest of foreign bankers' banks to behave in a way that would
amount to a timely redistribution of gold reserves. The Bank of France, with its strong
aversion to variations in its rate, had a particular motive for it. Whether the "help" was
extended to the Bank of England or to the market is immaterial. English paper was
actually taken, and gold released, by the Bank of France in 1906 and 1907, and this eased
the situation and prevented further rise in bank rate. Whether this was done by arrange-
ment or not is not particularly interesting.
2 The latter were probably more important than the former. Though the danger-
signal effect must be borne in mind, the Bank had, as we have seen, not much to do with
stopping normal booms. The classical theory of bank rate exaggerated the importance to
general business of moderate variations in rate as much as more modern theories do.
674 BUSINESS CYCLES
so many people failed to see the dependence of that effectiveness on the
historically unique technical position of the London market and, instead,
tried to explain it by a perfectly general — and unrealistic — theory. This
theory, on the one hand, greatly overstressed the influences which bank
rate can exert on foreign exchanges and gold movements through suc-
cessive effects on volume of domestic transactions, volume of deposits,
price level and incomes, balance of international commodity trade. On
the other hand, it never succeeded in properly defining the places that
open-market rates and reserve proportions hold relatively to each other
in the rational schema of central bank policy.
The presence of that light cavalry of English finance then explains
why both bank rate and gold movements proved, comparatively speaking,
so little disturbing to business and incidentally to the course of cyclical
phases. As regards foreign exhanges in particular, a cyclical regularity
was established by its action, which should be noted in passing. In the
absence of any chances, risks, and costs incident to the transfer of short
balances, the short-capital mechanism would obviously have tended to
equalize comparable open-market rates in different countries. As
between countries of different monetary standards, the element of risk
and chance was dominant. But as between gold-standard countries,
variations of exchange rates were confined within gold points. Hence,
it was possible for the operating banker to calculate maximum risk and
chance and to compare the result with the given difference between
comparable rates ruling in different countries.1 Given the responsiveness
of the short-capital mechanism to fractional gains, this would not only
limit possible deviations from equality of market rates,2 but also tend to
enforce parallelism of variations in those differences and variations in
rates of exchange. This expectation is disappointed as regards the rela-
tion between London and New York, but comes out very well as regards
the relation between London and the European gold-standard countries,
especially Germany and France.3 Still we cannot expect that parallelism
1 There were, however, several difficulties about doing this. We are assuming that
there were no risks other than those implied by the possible range of variation in rate of
exchange. If we take the highest grade bills only, this assumption works under normal
conditions. But it fails in any really serious crisis and also in the presence of political
fears. Moreover, gold points were somewhat variable, both in time and as between banks,
some of which may have had opportunities to ship more cheaply than others.
2 As far as the writer knows, this "solidarity" between the open markets of gold coun-
tries was first investigated by N. E. Weill, Die Solidaritat der Geldmarkte, 1903, though
the main facts about it were, of course, familiar to all writers on banking and finance.
3 See the standard work on the subject, H. Neisser, Der Internationale Geldmarkt vor
und nach dem Kriege, Weltwirtschaftliches Archiv, April 1929, especially charts on pp. 187
and 189, also, for what follows the chart on p. 221, which compares the curve of the distance
between the London-Berlin exchange and the difference of market rates with the Thomas
index of English business conditions.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 675
to be perfect. There are transactions that are less sensitive, or even not
at all sensitive, to those differences. We have to do, moreover, with a
surface mechanism which cannot always control the influence of the more
fundamental factors behind it. Finally, anticipations as to the behavior
of exchanges in the relevant future will work differently on different
levels as well as in different seasonal and cyclical situations. It has, in
fact, been shown by Professor Neisser that the graph of the distance
between the curves of exchange rates and of the difference between market
rates displays fairly regular cyclical properties, which are what emerges
behind the apparently erratic behavior of exchange rates considered by
themselves. It does not follow that we should expect any consistent
short-run relation between foreign exchanges and domestic market rates
alone, or that we should be able to predict any definite effect from a given
absolute change in domestic bank rate alone, however completely the
Bank may succeed in making it effective. The opinion which it is said
prevailed in the city and in support of which Goschen's authority might
be invoked, viz., that an increase of 1 per cent in bank rate would, and
a smaller increase would not, turn unfavorable exchanges, was therefore
not exact.
3. In itself, the mere fact that there was in the world of prewar
capitalism one gold market which in importance and accessibility over-
shadowed all others1 would not greatly matter for our purpose. It would
merely offer an occasion for the remark that, first, because of the "autono-
mous" movement of gold from the countries which produced it, and
second, because of the special transactions effected by central banks or
other monetary authorities, no expectation as to cyclical behavior can be
formed with any confidence. The cyclical migrations, which the units
of a constant stock of gold could have been expected to display under a
regime of unfettered gold currency and in the absence of any regulatory
interference, are too clear to detain us.2 But all the more relevant is the
further fact that the gold market, toward which, with the partial excep-
tion of the production of the United States and Russia, practically all
newly produced gold gravitated, was located in the same country and
controlled by the same financial organism that also controlled the inter-
national short-loan fund. Because of this, England was in a position, by
extending or restricting short credits, to enable operators in other coun-
tries to buy gold or else to prevent them from doing so, not absolutely
1 There were many local ones all over the world. In Paris a price was regularly quoted
for bars which was distinct from the gold premium charged by the Bank and from the price
of gold exchange.
2 This does not mean that they do not offer any problems, but only that the problems
relevant to our argument can readily be solved if we amend the classical theory of gold
movements by a theory of the short-balances mechanism and stress differences in interest
rates rather than differences in pricelevels.
676 BUSINESS CYCLES
but to a considerable extent. As far as the Bank of England controlled
the domestic open market, it may, therefore, be said to have controlled
international gold movements and to have — indirectly — acted as the
bankers' bank of bankers' banks. This strengthened its position and its
ability to mitigate the impact of foreign disturbances on domestic busi-
ness situations in a way not open to any other central banks. The
Reichsbank, the Bank of France, and other central institutions aimed at
similar freedom of action by other means. Some of v these were also
adopted by the Bank of England. It was, in fact, a very natural dis-
covery to make, that actual or expected gold movements might be
influenced directly and that this method would serve to handle a class of
short -run difficulties as effectively as measures which, besides doing this,
would also produce undesired effects. But while other central banks
began at the turn of the century to head toward the gold-exchange stand-
ard, which meant, though they may not have been aware of it, moving
away from gold-standard ideas, the Bank of England, up to the war, never
did anything1 that pointed in that direction. All it attempted to do was
to smooth the working of the unfettered gold standard, which for the
time being was clearly in the interest of England and which remained her
long-run policy. It, therefore, confined itself mainly to acting on gold
points by varying its price for bars and foreign coins, refusing to sell bars,
giving free advances on gold imports, and so on, and in all normal cases
continued to rely on its influence on money rates. To insulate the
domestic price level was no part of this policy. To stabilize it was part
of the policy only in the sense that — never mind what the Bank's leading
men may have thought or said — it tended to mitigate temporarily the
impact of noncyclical factors.
In the late eighties and the early nineties — from 1879 to 1888, total
coin and bullion in the Bank of England fell (with fluctuations) by about
one-third, while total deposits in the United Kingdom rose by, roughly,
one-third — the problem was one of "scarcity." It was during that time,
as far as the writer knows, that the Bank first resorted to gold policies,
which, however, proved particularly effective during the troubles of the
early nineties.2 What would have happened if no increase in gold pro-
duction had occurred is easy to guess: new methods of economizing gold
would have been resorted to. As it was, South African gold reversed the
situation. Its modus operandi is highly instructive. The Bank at first
simply discontinued its gold policy and allowed the new gold to accumu-
late in its vaults — its gold stock increased rapidly from 1891 to 1895, and
1 Fulfilling its legal obligation to sell sovereigns in minimum-weight coins was as near
as it went to refusing to play the orthodox game.
2 In 1893, when the inflow of gold had already set in, the price of gold in the London
market rose to 8 pounds, 17 shillings, 10.8 pence per ounce standard, while bank rate
stopped at 4 per cent.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 677
its reserve proportion moved, substantially, parallel to it1 — or to flow
through the London market into other banks2 and to other countries,
while the price level continued to fall. The effect on open -market rates is
as well marked as we should expect, and was temporarily intensified by
the payment to Japan of the Chinese indemnity, which involved the
transfer to London in October and November 1895 of 13 million pounds
raised by the Chinese government in Paris. But it should be obvious
that even this effect of the new gold cannot be interpreted in terms of
monetary factors alone and that it was so much in evidence only because
of the cyclical situations on which the new gold happened to impinge.
For, in the prosperity that followed, it entirely disappeared, and rates
stiffened in spite of the swelling tide of gold production. Prices, on the
other hand, then began to rise. They probably rose more than they
would have done in the absence of such a tide, exactly as money rates
had no doubt fallen more than they would have done in the absence of it.
Neither effect is being denied here, but neither was strong enough to
assert itself except in cyclical phases favorable to it. The writer does not
see that this interpretation leaves any major fact unaccounted for or
that there is in the sequence of these events anything to contradict our
views about the cyclical process, its relation to monetary factors, and the
facts and possibilities of central bank policy.
To complete this part of our argument, we will notice that even before
prosperity (in our sense) had set in and while revival (in our sense) was
in full swing, with low rates, low prices, and great activity in building
and in trade in general, the American troubles of 1896 and the incident
drain of gold drove the Bank back upon raising its selling price for bars.
It also raised its rate, which, in a very liquid market, it proved extremely
difficult to make effective. The latter measure, not easy to understand
from the situation in which it was taken, was ratified by events, for that
1 See W. E. Beach, British International Gold Movements and Banking Policy, 1881-
1913, 1935, p. 71.
2 This was then, but especially later, facilitated by what may be called the Gold Scare.
Ever since the publication of Bagehot's Lombard Street, financial writers had been preach-
ing that the "gold basis" of the British banking system was too small. These preachings
began to take effect in the late nineties, and the great London banks used the opportunity
afforded by the new plethora of gold to build up gold reserves of their own. Whatever
we may think of the sentiments voiced in Sir Felix Schuster's addresses, for example, and
of the reality of the "dangers" incident to smallness of the "ultimate reserve," the scare
certainly helped to produce, together with the demand for gold from the countries which
at that time went on the gold standard or made preparations for doing so or simply wished
to broaden the basis of an already existing gold currency, much the same results as could
have been expected from an international policy of partial sterilization of the new gold.
This is one of the reasons why it is not easy, on the one hand, to interpret the gold move-
ments of that period and, on the other hand, to estimate the net effect of the new gold on
price levels.
678 BUSINESS CYCLES
situation shaded off into a burst of prosperity the following year when
the price of bars was raised to £ 3-18-J^, but bank rate to not more than
3 per cent. The course of events in America and Germany, the require-
ments of foreign financing and boom conditions at home, account for the
use of gold devices along with bankrate in the next two years, during
which the Bank's gold stock fell to the level of 1894. Again, American
troubles and war finance account for their use instead of additional
manipulations of bank rate, from 1900 on. As the Kendratieff wore on
and gold continued to flow in, the Bank's position grew stronger of
itself; and toward the end of the period, protective gold policies fell
almost into disuse. There are, in fact, traces of a preoccupation with the
opposite problem, i.e., the problem of preventing the gold stock of the
Bank from increasing to amounts the very sight of which would have
encouraged excesses. It was, however, kept, in its yearly averages, at or
near the 30 million level until 1913.
No such gold management was, of course, possible in the United
States, where, in consequence, we find during the sixteen years preceding
the war much more parallelism between national bank deposits and the
total monetary gold stock. The German case is for our purpose less
interesting. The Reichsbank was mainly concerned with strengthening
the bases of the German standard. It was a steady buyer of gold from
the first, acquiring over 434 million dollars of it between 1876 and 1893.
In the times of prosperity that followed, such gains as were made were
more than absorbed by circulation. But from 1907 on, the accumulation
of a larger reserve tended to become an end in itself, and all other con-
siderations were subordinated to it. By purchases and the use of gold-
exchange standard and other devices, among which the substitution of
the Reichskassenscheine for gold in circulation may be mentioned, it suc-
ceeded in acquiring the ample gold outfit with which it entered the war.
D. Stock Exchange Series. — These series or phenomena described by
them have been dealt with at various stages of our argument. We
return to the subject merely in order to round it off, particularly with
reference to banking and bankers' banks' policy. For this purpose, it is
convenient to define the stock exchange as that market which deals in
bonds and shares. Another part of the market of "capital values," the
realty market, will not be considered here in spite of its considerable
cyclical importance.
Although the stock exchange is, as has been stated in Chap. XII,
really a part of what in the widest sense we have called the open market,
it will now be looked upon as distinct from it, though communicating
with it. We know where the money comes from that buys the various
"branded articles" that constitute the "commodities" of this very
imperfect market. It comes, first, from the "surplus funds" of banks,
THE CENTRAL MARKET AND THE STOCK EXCHANGE 679
lending to the stock exchange being the most important method of banks'
temporary investment. Second, it comes from nonbanks (although often
through bank loans on account of others1), domestic business concerns,
and foreign banks and capitalists, who also want to invest funds tem-
porarily— the latter source sometimes enabling member banks to thwart
bankers* banks' policies, and the stock exchange to thwart member
banks. Third, it comes from "investors" — from the savings of house-
holds and from the funds of nonbank firms and of banks which also may
wish to invest permanently. Finally, banks directly finance trans-
actions of both speculation and investment by collateral loans to cus-
tomers, which, however, must not in general be identified with financing
speculation or households' investments although we now shall tentatively
assume that this is admissible in the case of the collateral loans of New
York banks. Though all banks lend for stock exchange purposes, their
relation to and position in the stock exchange varies, as we have seen,
very much in time and as between countries, and this difference does not
disappear if we include among banking firms, as we really ought to,
brokers and jobbers or Paris agents de change. They still essentially
remain middlemen and never have the position of German banks. The
difference between a speculator and an investor can be defined by the
presence or absence of the intention to "trade," i.e., to realize profits
from the fluctuations in security prices. But since investors also bor-
row and since they may at any time turn into speculators, this does not
overcome the difficulty of distinguishing speculative from nonspeculative
transactions. More useful for practical purposes is the criterion of the
margin account.
1. The shifting of old issues between speculators, investors, and
banks, for a variety of purposes with which we are familiar and which
form an essential link in the mechanism of credit creation, constitutes
one group, the placing of new issues with speculators and investors,
another group of transactions relevant for our purposes. Disregarding,
for the moment, the latter group, we will first consider the role in the
cycle of transactions in "old" stocks and bonds and of the financing of
these transactions. Needless to say, the phenomenon would, in the
absence of external disturbances, owe its existence entirely to the cycle,
and must obviously be expected to be one of the most regular of its fea-
tures, so much so that it would be hard to find instances of cycles that
would not display it. It is clear, moreover, that the dissaving implied in
living on gains from speculative or nonspeculative transactions of that
kind2 has much to do with the amount of consumers' spending in, and
1 The role of Temporary Investment and, in particular, of Brokers' Loans will be dealt
with in Chap. XIV, Sec. F.
2 The gains of speculators include, no doubt, an operational element which is not a true
capital gain but of much the same nature as any other return to personal activity — for
680 BUSINESS CYCLES
with the psychic atmosphere of, prosperity, and that, similarly, losses in
recession or depression have much to do with the reduction of consumers'
spending in, and with the psychic atmosphere of, recession or depression.
Nor need we stay to show how increases and decreases in the value of
collateral may be responsible for loans (also for business purposes) which
would not have been granted or would not have become endangered
without them. As everybody knows, this may be an important element
in the banking situation of "crises*5 and in vicious spirals in general.
Peculiarities of the pricing process on the stock exchange and the
effects of speculation on the situation of member and bankers' banks and
on the financing of business are conveniently dealt with by reference to
the time-honored question as to whether "the stock exchange absorbs
credit."1 The naive argument in favor of an affirmative answer has been
met, and for practical purposes can be met, by pointing out that, norm-
ally though not in times of speculative manias, "funds" lent on the stock
exchange would, to a great extent, be otherwise unemployed. Our
previous analysis of the behavior of banks in cycles lends some support to
this. The fact that in the presence of high renewal rates, banks and other
lenders turn from the commercial-paper market to the stock market does
not substantially invalidate that reply, nor does the further fact — the
importance of which greatly varies according to the technique of specula-
tion in different countries — that anyone who wants to speculate must
"buy himself in," which may imply withdrawal of existing balances or
facilities of creating them from other uses, without necessarily releasing an
equal amount at once.
More fundamental is it, however, that rising stock prices are not in
the same way associated with increasing use of balances as are rising
prices of ordinary commodities. Obviation plays a greater role2 than
it can play in the latter case, in New York, particularly, since the founda-
tion in 1919 of the Stock Clearing Corporation. Moreover, brokers'
loans induce no cashing of checks and, consequently, no loss of cash to
circulation. Again, most of that portion of brokers' deposits that is
derived from l&ans need not actually appear as deposits at all, because it
is entered directly under the banks' certified checking accounts, against
instance, the professional income of a lawyer. Spending this on consumers' goods is not
dissaving. There is no need, however, to emphasize that element here.
1 There is a large literature on this question. We will only mention Professor Mach-
lup's Borsenkredit, Industriekredit und Kapitalbildung, 1931; R. N. Owens and C. O.
Hardy, Interest Rates and Stock Speculation, 1925; Thomas Balogh, Absorption of Credit
by the Stock Exchange, American Economic Review for December 1930; an interesting dis-
cussion by Professors Cassel, Spiethoff, and Hahn in the Frankfurter Zeitung in May 1927;
and R. Reisch, Riickwirkungen der Borsenspekulation auf den Kreditmarkt, Zeitschrift fur
Nationalokonomie, 1929. The question will be taken up again in the next chapter, sec. F,
2 See Professor Machlup, op. cit.> p. 79, et seq.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 681
which no reserve is required.1 These statements may, with proper
modifications, be paralleled for European stock exchanges.
But the essential point, which really arises out of the same circum-
stances and also helps to account for obviation's being so easy in this case,
is this. What we have termed the efficiency of money bears a rela-
tion, fundamental to any theory of money, to institutional periods of
payment, which, in turn, bear a relation to the timing and periodicities of
economic processes. There is no such time element in transactions on the
stock exchange, because there is no process to go through. Institutional
arrangements and the limits of physical possibility still produce a mini-
mum period, but it might conceivably be one of minutes only. Hence,
there is almost no limit to a "price level" that a given "fund/* however
small, may "support." Moreover, while a commodity must, at the
price determined in its market, move over its track through the economic
organism, there is no such necessity for stocks. If, in the case of a
commodity, market price is such as to prevent it from starting on, or com-
pleting, its course, stocks of the commodity accumulate, which in them-
selves are irksome maladjustments and must eventually be liquidated.
This keeps each price ultimately in line with other prices and in subjection
to the monetary ligarnen. But in the case of, say, a common stock, the
fact that at its price it cannot "move" docs not necessarily spell an
untenable situation. The people whose action and estimation is responsi-
ble for that price may be willing to keep it, and normally can do so.
The stock need not move to any definite spot in the economic organism, in
order to fill its economic function. And the mere fact that a small group
of people entertain an irrationally high opinion about the value of a stock
can, with very small transactions, raise it to indefinite heights without
either absorbing any funds or credit facilities or eliciting pressure from
monetary magnitudes. This is the great difference between pricing on
stock and pricing on produce exchanges.2 Technical qualifications apart,
into which we need not enter, such as the holding of war chests by pools
and individuals, it is, therefore, true that stock speculation docs not
absorb money or capital or credit in the sense usually implied by this
phrase.3 New issues, of course, do "absorb funds," but only to release
them, unless the proceeds be applied to the repayment of bank loans.
1 H. Rogers, Effect of Stock Speculation on the New York Money Market, Quarterly
Journal of Economics, vol. XL, p. 453.
2 It has been pointed out that a similar peculiarity exists in many cases of commodities
in the proper sense of the word, such as pictures by old masters and real estate. This is
quite true, even though certain other characteristics of the stock market are absent. But
this fact does not constitute an objection — it rather illustrates our meaning.
3 This view, together with some of the propositions that are to follow, is often taken to
imply "friendliness" to, or an intention to "defend," stock speculation. In order to dispel
any such suspicion, the writer begs leave to state that on moral, political and cultural
BUSINESS CYCLES
2. In analyzing the pricing process on the stock exchanges, another
consequence of the fact that the "commodities" of this market are, by
virtue of their economic nature, held rather than moved must be taken
into account. A market in which this is the case and in which the pres-
sure of the monetary ligamen is not present in the same sense as it is in
markets of ordinary commodities, in which, moreover, stocks are to a
great extent held for a rise but, serving as collateral, have to be liquidated
in case of a fall in price, in which " supply " will normally^ increase in
consequence of a fall, decrease in consequence of a rise, or also vary
without any variation in price — such a market is not likely to conform to
traditional ideas of the behavior of supply and demand. If it does not
behave more erratically than it does, this is due to the fact, easy to verify,
that the varied responses which a change in the situation elicits in part
balance each other. The resulting tone or tendency is, as a rule, so
clearly marked, because it is, as soon as discerned, acted upon by specula-
tors. In a sense this is true of any market; but for the reasons stated
above, it holds for no other market to the same extent.
Therefore, explanation of that tone or tendency has been rightly felt
to be a distinct task, i.e.9 a task not only distinct from the explanation of
market phenomena in general but also of the phenomena in other specu-
lative markets. Speculation in raw material futures is, for instance,
much more likely than speculation in common stock to display those
features that can be listed under the traditional and self-explanatory
headings; insurance, arbitrage, and steadying price over time. Carrying
temporarily stock that people do not, or do not yet, care to carry perma-
nently, and sending monetary capital into the firing line come nearer to a
description of what speculation in stocks actually does.1 What follows
refers primarily to the behavior of stock prices. The behavior of bond
prices has been dealt with in the preceding chapter. But it must be
remembered that the pricing of low-grade bonds involves a speculative
element, which will tend to make their prices behave much like prices of
stocks.
Since stock prices have more degrees of freedom than other prices
have, and since financial groups — pools and others — confront a public
very much more excitable and very much less intelligent than the constit-
uent individuals are in their ordinary business pursuits, it is tempting to
stress mere mass psychology, on the one hand, and mere abundance or
grounds he personally welcomes almost any measure in any way hostile to it, regardless of
economic consequences. In particular, he would welcome an enactment making specula-
tion a misdemeanor for the members of certain professions.
1 The theory that no concern is a permanent source of net revenue allocates, however,
as will be clear on reflection, to trading in stocks a significance within the capitalist process
which is additional to the "functions" of it that enter into traditional formulations.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 683
scarcity of funds, on the other. While there cannot be any doubt about
the presence of either element, neither of them makes, by itself or jointly
with the other, a satisfactory explanation. As regards the first, there is,
of course, much more scope for waves of optimism and pessimism on the
stock exchange than there is in industrial and commercial business, and
it is more nearly true for the former than it is for the latter that anticipa-
tions may temporarily produce the expected result. But even manias
have their starters, and we must not judge by manias only. It should be
added that there has been a slow but persistent change in the behavior of
the public, which becomes obvious if we distinguish irrational optimism
as to general conditions and irrationality in the choice of the particular
stock to speculate in. The scene of 1929, perhaps, differed but little
from the scene of 18731 or even from that of the South Sea Bubble, in
the former respect; but it differed significantly in the latter. In order
to choose rationally, a man need not be able to analyze a business situ-
ation or the state and prospects of an industry or concern. All he needs
to know is what advice or example or current opinion to follow. It
remains true that irrational fancy and downright foolish hopes or fears
count for much in the short run. But it is no less true that they never
provide the motive power of a boom and that they never prevent the
real state of things from asserting itself eventually.2 This real, or objec-
tive, state may be decomposed into external factors, cyclical phases, and
conditions peculiar to individual industries and concerns. No doubt,
the word expected may usefully be inserted in each case.
While the above analysis of the peculiarities of stock pricing thus
gives some, if limited, support to an optimism -pessimism theory, it
rather suggests that less emphasis should be placed on money-market
factors than we should infer from the familiar statistical evidence. We
have seen in the preceding chapter that small variations in open-market
rates made a great deal of difference to the operator in bonds and hence
to their prices, but no variations of rates such as currently occur in a
normal cycle can possibly matter for speculation in stock. It is, there-
fore, availability rather than cost of credit that we should look to. But
we have seen that, barring new issues, comparatively small amounts will
go a long way on the stock exchange. Considerable booms can develop
1 Even that is not quite certain, however. Foreign buyers in 1929 need not have been
unduly optimistic as to American conditions in order to have preferred American stocks to
European ones. And if American holders of American stocks failed to sell in time, this
does not prove that they foresaw nothing but continuing booms, but only that they did not
foresee that the loss from a fall in security prices would be greater than the amount of
income tax which would have been due had they sold out.
8 The stock exchange thus affords an excellent example by which to demonstrate what
"psychological factors" and, in particular, "expectations" can and cannot do and in what
sense economic reality is independent of them.
684 BUSINESS CYCLES
on a narrow basis of "cash," and unless banks set their faces with what
would be quite unusual determination against the speculative purpose, it
is not easy to see how speculation could be starved by lack of funds,
though, of course, new issues could. Hence, we shall be inclined to
attribute that statistical regularity more to coincidence within the cyclical
process — to be explained presently — than to the causative influence of
monetary factors, for example, in the sense of the theory according to
which funds migrate from industry and commerce into the stock exchange,
thereby igniting speculative booms, and then back again, thereby putting
a stop to then.
3. Indices of stock prices make synthetic and systematic series
descriptive of a primary and consequential phenomenon. These series
will, of course, be strongly cyclical, however they may be constructed.
But our model yields no generalization about trend. Such trends as we find
simply reflect historical facts, which must be separately accounted for in
each case, and depend on the properties of the index chosen. If, for
example, the index expressed average price per dollar of paid-up capital
plus accumulation for a number of identical concerns, it would, barring
changes in the purchasing power of the monetary unit and in the value of
natural agents owned by those concerns, normally go to zero after a
period of sufficient duration. If new concerns are currently added and
old ones currently dropped, it will not do that, but still need not display
any particular result trend in our sense.
Since we do not believe in the dominant role of money-market condi-
tions we shall not expect any strong seasonal variation, although holidays
and so on must exert some influence. As regards cyclical variation, three
facts must be borne in mind. First, the public's anticipations concerning
cyclical phase and its anticipations concerning the fortunes of individual
industries and concerns are interdependent and tend, within the short or
medium run, relevant to the decisions of speculators and even investors,
to point in the same direction. Even an old and ailing concern will in
the majority of cases do better in prosperity than in depression. But
both speculators and investors are particularly attracted by relatively
new — though perhaps not so much by quite new and untried — industries
and by the leaders in those industries which carry prosperities. That this
is so is clear — railroad, electricity, motor, rubber, oil stocks and dozens
of equally familiar instances lie at hand to verify it. But it should be
noticed how well this fact illustrates the "function" and also the "ration-
ality" of speculation; for the diagnoses which this behavior implies are in
general well verified by objective events, themselves independent of those
anticipations. Not, of course, in all individual cases, nor necessarily from
the standpoint of the speculator — but even when the latter loses his
money, he rarely, if ever, acts on dreams: the developments he counts
THE CENTRAL MARKET AND THE STOCK EXCHANGE 685
upon almost invariably mature, however seriously he may be at fault as
to timing.
Second, the stock market is imperfect, but it is freer from friction
than almost any other. Decision can be given effect to and reversed
with a promptness altogether impossible in industry. Depressive
situations may so weaken hands as to force them to relax their grip, even
after opinions have undergone a favorable change, but on the other hand,
the effects of such a change may be accentuated by the supply being
suddenly curtailed when demand revives. Hence, it is natural to expect
that upward movements on the stock exchange will, in general and in the
absence of unfavorable external factors, set in earlier and gather force
more quickly than the corresponding upward movements in business,
i.e., often come about already in the later stages of revival when things are
beginning to look better every day, with new possibilities showing them-
selves. Similarly, it is to be expected that stock prices will turn before
other indicators, i.e., when in the later stages of prosperity limitations
and difficulties emerge and it becomes clear that possible achievements
have been fully discounted.1 Now, in the first situation, interest is low
and the money and open markets are normally easy; in the second,
interest is high and the money and open markets are normally tight.
Some causal significance may be attributed to this relation, particularly
if we include the role that monetary conditions will play in the prognosis
by speculators or their advisers of what is ahead in business.2 But the
principal explanation is that both stock prices and money rates respond
to cyclical phases so strongly that their variations would be regularly
associated even if they had nothing to do with each other.
Third, because of that promptness in reaction and also because the
self-reenforcing mechanism is so strong in the stock market, we must
expect movements to outrun their momentary goals much more than in
other markets and reactions to be correspondingly sharp. Even if the
market be undecided, this will mean fluctuations — of the sort we called
hesitations in Chap. IV, Sec. E — rather than inaction. Moreover, while
general business may and often does settle down into recession in a per-
fectly orderly way, this is hardly imaginable in the case of the stock
exchange. Recession means reduced profits and, for many concerns,
more or less serious troubles. It gives scope to bear attacks. But even
1 That view is supported by the observation that very frequently the rise in the leading
stocks stops before the top of the market is reached. Speculative interest, realizing that
their possibilities have for the time being been exhausted, looks around for stragglers that
may still offer secondary chances. That is when the "froth" comes up.
2 The writer is inclined to believe that the facts that the weekly statements of the New
York Associated Banks used to be a feature in the Saturday market and that similar atten-
tion was in London and Berlin paid to the statements of the Bank of England and of the
Reichsbank, were entirely due to this.
680 BUSINESS CYCLES
if nothing of this sort ever happened or were anticipated, the mere fact
that there is no reason to expect, save in particular cases, any upward
movement would suffice to make speculators lose interest in their hold-
ings. Thus the upper turning point or the approach of it is, in this case,
likely to mean a crash, which need not greatly affect the general business
situation. But although it need not, it often will, both by upsetting the
credit situation and by uncovering weak spots. It is easy to carry this
argument through spirals and depressions into recovery and to satisfy
oneself that in the absence of an abnormal amount of reckless finance and
misconduct and often in spite of their presence, bear markets do not last
unless coinciding with and independently induced by, depressions in
general business. This is perfectly compatible with the influence, pre-
viously recognized, which stock exchange events exert on consumers*
spending.
For corroboration we return to Chart XXXVI. The fairly marked
covariation of railroad stock prices — as long as these were the dominant
element, i.e., to the nineties — and of railroad and industrial stock prices
with New York loans and deposits, and their strongly marked covariation
with New York Clearings1 should again be noticed first. In both railroad
and industrial stock prices the Kondratieff prosperity from 1897 on
shows well and so do the Kitchins. The major movements which we
observe, however, clearly reflect the Juglars: we see the (anticipating)
boom of 1868 and 1869 and then the characteristic slump from 1873 to
1877; then the (also anticipating) boom from 1877 to 1881; the same
phenomenon, most regularly repeated from 1885 on; no such precedence
for the first Juglar of the third Kondratieff, which may have been due to
the aftereffects of 1893 and political factors; but more regular behavior
again in the second. This reflects the relation, which according to our
analysis should exist between speculation and the investment process.
The latter, as we know, dominates the Juglars more obviously than the
Kitchins. All series that link up with investment, such as pig-iron
production, employment, and so on, display Juglars with particular
emphasis. The fact that stock speculation behaves similarly is, hence,
of some significance. The reader should again inspect Chart XXXIV,
particularly the graph of New Listings, for which speculation paves the
way.
Charts XXXVII and XXXVIII substantially tell the same tale. In
the first there should be noted the instructive differences that comparison
with the American series reveals, among them the spurt in stock prices
from 1895 to 1897, strongly underlined by the behavior of call rate, after
which they fluctuated around a declining level. The second presents a
1 See A. M. Matthews, New York Bank Clearings and Stock Prices, 1866-1914, Review
of Economic Statistics for October 1926, where trend ratios are compared.
THE CENTRAL MARKET AND THE STOCK EXCHANGE 687
picture which is much more, though not quite, similar to the American.
German stock prices reproduce the movement of American industrials up
to, and down from, the peak of 1872. In reality, however, this veils a
difference in behavior, since the German index includes all leaders, while
the American index excludes railroad stocks, which in this country, more
agreeably to expectation reached their peak sooner, i.e., in 1869. Political
events — the Franco-Prussian War — may plausibly account for this differ-
ence. The trough occurs in the same year as in this country, 1877,
but the subsequent rise already stopped in 1880, the rise of 1888 in
1870
1875
I860 1885 1890 1895 1900
CHART XXXVII— (See Appendix, p. 1064.)
1905
1910
1914
1889, after which the index falls, preceding the American index to the
end of 1892. Our guess that political events interfered with the course
of things after 1893, this time in America, is supported by the fact that
in Germany a strong upward movement, which lasted through 1898, had
already begun in 1894. This would make a very normal case, not at all
impaired by the duration of the interval — over three years — by which the
stock exchange shadow preceded the form that cast it, viz., the prosperity
which we date from 1898. We need only look at the general situation,
which may be described as buoyant revival, and at the stock exchange
leaders: everything was in fact "looking better every day" and the con-
tours of the achievements in the electrical field unmistakably stood up
before everyone's eyes. By the end of 1898, on the other hand, it
was not less clear that possibilities were for the moment being rapidly
exhausted — more than exhausted — and that the immediate future offered
little for the fancy of speculation to feed on.
688
BUSINESS CYCLES
THE CENTRAL MARKET AND THE STOCK EXCHANGE 689
No undue amount of conscious rationality is by this interpretation
attributed to the rank and file of speculators. But it is held that objec-
tively there was method in the madness. In this respect it is also worth
while observing how close stock prices come to reproducing the move-
ment of dividends.1 And both go fairly well with Foundations and
Issues of Industrial Stock. While it is unnecessary to dwell on the
significance of this fact, a few minor points should be noticed. Founda-
tions and Issues do not display as close a covariation as one might expect;
on reflection, however, it will be realized that the former, being of the
essence of industrial booms, will cluster at or before the peaks of business
activity, while the latter, being tied to the opportunities created by
speculation, will cluster at or before the peaks of stock exchange activity.
Issues of industrial bonds and issues of stocks should display similar
" trends" but opposite fluctuations. The former expectation is verified,
but the relation between the fluctuations is less consistent than we might
have hoped. A peak occurs in both series in 1911. Juglars show excel-
lently in stock issues, Kitchins rather better in bond issues.
4. In this and in earlier chapters facts and arguments have been pre-
sented which suffice to draw the contours of a theory of the relation of
banks to stock speculation and to that part of the investment process of
which the stock exchange is the scene. We may sum up by stating that
even where banks most strictly keep to the principles of classic "deposit
banking," considerations arising out of the practice of lending, to the
stock exchange and to customers, on stock exchange securities provide a
most important link between the prices of the latter and the fortunes of
each individual bank; but that where banking of the credit mobilier type
prevails and banks directly "patronize" industries, control their financial
policy, manage their issues, hold stocks or bonds and trade in them on
their own account, the stock exchange may become the pivot of their
business. But we will now confine ourselves to the relation of bankers'
banks to the stock exchange.
From the proposition that speculation in stocks does not, or not to
a significant extent, "absorb credit" — that the stock exchange is not a
sponge but a channel — it does not follow that it is a matter of indifference
to the central bank. It is obvious that new issues are not. For they
precisely supply what it is the function of a central bank to regulate.
They may increase the lending power not only of individual member
banks but of the banking system as a whole. As we have previously
seen, they may, if for foreign account or merely if internationally financed,
1 Prices of the stocks of individual concerns are also — everywhere — closely associated,
though only in the short run, with the prices of the products of those concerns. It would
be dangerous to stress this relation too strongly. It points, nevertheless, to the influence
of obvious common-sense considerations on the decisions of speculators. The relation
is instantaneous, however, and this type of decision is likely to be taken postfestum,
690 BUSINESS CYCLES
upset the working of the mechanism of short balances and affect gold
movements and the central banks' reserve. Barring this, they always
affect the open-market situation. And they are associated with reckless
financing. It stands to reason that the central bank cannot be indifferent
to that.
But if central banks cannot be indifferent to the volume and the pur-
poses of new issues, they cannot be indifferent to the movement of stock
prices, because a boom in these is, practically speaking, the necessary
and (almost) the sufficient condition for a boom in issues. Moreover,
even trading in old stocks influences business activity — the intensity of
the Secondary Wave — through the withdrawal and the spending of
speculative gains, and also the general banking situation, especially if
member banks themselves are sellers or buyers on their own account.
This need not cross and may even help the policy of the bankers' bank,
as we have seen. We have, in particular, had occasion to observe that
the selling by banks of stocks and bonds may effectively restrain a boom
and that their buying may mitigate a slump. But this need not be so.
It follows that the untenable theory about speculation absorbing funds
may be, after all, the ill-tailored coat of much practical wisdom : though
for a different and, in fact, almost opposite reason, there are plenty of
motives for bankers' banks to react to stock speculation and to try to
fight its excesses during prosperity or the later stages of recovery, and
vice versa bear attacks during recession and depression. Success in this
may avoid many hitches in the investment process, reduce many plus or
minus deviations of current business, remove some of the darkest hues
from the picture of "deep" depressions.
The partial analogy that exists between new issues and open-market
operations by the bankers' bank suggests that the former may, to some
extent at least, be controlled by the latter or, more generally, be regulated
by a central bank that keeps a firm hold on the open market. New issues
are, in fact, the most important link between stock exchange booms and
easy money, and sensitive to tightness. Moreover, there are other and
more direct means to the same end which it is and was within the power
of, say, the Reichsbank or the Bank of England to use. But acting on
the conditions that facilitate or impede new issues is a delicate task. It
might prevent financial consolidation in individual points of potential
danger or even bar the way to consolidation of an overwrought situation
in the banking system and thus precipitate what it is the intention to
avoid. In fact it is impossible as a general policy. Some effects on the
economic process are, moreover, as we have seen, produced by booms in
the prices of existing stock, independently of whether or not they induce
new issues. It has, therefore, been at all times the almost universal
opinion of practitioners of central banking that the management of
THE CENTRAL, MARKET AND THE STOCK EXCHANGE 691
credit would never be fully effective and especially never effective in time,
unless stock speculation be acted on directly.
Nevertheless, we see throughout the period under survey very little
effort to accomplish this and very poor results of such efforts as were
occasionally made. As we have seen in Chap. VII, something was done in
Germany, but it was done by legislation and not by central bank action.
The reason for this follows from our analysis of pricing in the stock mar-
ket. Since stock speculation does not absorb funds, it must be extremely
difficult to stop or to restrain by any of the ordinary tools of central bank-
ing. Bank rate in particular is anything but omnipotent with respect to
production and commerce, but it is, within the limits that are practically
feasible, almost ineffective with respect to stock speculation: regulative
influence on stock prices presents a problem entirely different from that of
influencing commodity prices.
CHAPTER XIV
1919-1929
A. Postwar Events and Postwar Problems. — The formidable task of
interpreting, economically and sociologically, our own time cannot be
attacked in this book. Still less is it within our plan to recommend or to
criticize remedial policies or fundamental reforms, or even to enter into
discussion of individual measures taken or proposed. Whatever of this
the reader may find, explicitly or implicitly, in the following pages, is
incidental to an argument the very restricted purpose of which should be
borne in mind throughout. That purpose is to answer the question how
far the cyclical process of capitalist evolution, as analyzed for the 130
years that preceded the World War, can be proved to have persisted in
the postwar period, and to see how our model works under the conditions
and with the richer material of that period. The contribution toward
an understanding of the postwar-world which an investigation of this
kind can be expected to make may prove worse than valueless, if its
character, methodological background, and analytic intention are allowed
to drop out of sight. Wherever it seemed possible, an attempt has been
made to save space and to rely on the fact that current economic events
are, and have been since the war, very much more efficiently reported
than before, and on the hope that general contour lines are, therefore,
familiar to the reader. * It will be convenient to deal first with the years
preceding the world crisis and to comment on the latter in a separate
chapter.
We exclude the years 1914 to 1918— for Germany also 1918 to 1923
— on the ground that those years were dominated by "external factors"
to an extent that makes their figures valueless for our purpose. This is,
indeed, not quite true. The rhythm of economic life clearly persisted in
the United States, and some aspects of war events are not without rele-
1 It is, hence, suggested that the reader refresh his memory by going over the charts and
comments of one of the well-known services. The Harvard Service's quarterly surveys
are particularly recommended (for Germany, the surveys of the Institut fiir Konjunktur-
forschung, for England those of the London and Cambridge Economic Service). For
students of economics there is no more stimulating exercise than to bring to bear their
analytic apparatus on the task of interpreting the course of postwar events as described
in those reports. The most convenient basis for a study on a world-wide scale is, per-
haps, the material of the League of Nations.
602
1919-1929
693
DEPOSITS Pit
CIRCULATION
.US
J
/\ INTEREST RATE ON
'•PRIME COMMERCIAL FftPER
\ f\
\f '
V '.
ft a
vance for the study of business cycles. In particular, war expenditure
affords as good experimental evidence as we can ever hope to get
about the nature and consequences of
a boom which has nothing to do
with innovation and is brought about
by expanding credit and stimulating
consumption alone. The fact that
expenditure was not directed into
channels which would commend them-
selves to advocates of such a policy is
entirely irrelevant,1 for all that
matters is that depressions were actu-
ally impending or in progress in 1914
and that public expenditure turned
them into prosperity first and created
untenable situations afterward. But
although the case almost ideally com-
plements and illustrates part of the
argument of this book, we will follow
the practice of the majority of stud-
ents and eliminate those violent " irreg-
ularities" by leaving out the figures
of those years.
It is obvious, however, that exter-
nal factors in our sense continued to
play a supernormally important r61e
throughout the postwar period. That
our second component of economic
change, the cyclical process of evolu-
tion, was still present and asserted
itself in the same manner as before
is not so obvious. Owing to the his-
I I I I
I I I
I I I I
torical character of our subject — or l919 2^ 193° 3*
the fact that it is "institutionally CHART XXXIX.-United States postwar
,.,. ,,, ., . ,. , , ' pulse (see Appendix, p. 1065).
conditioned — this question would
arise in any case, even if there had been no war: whenever we wish to
apply our analysis to an additional span of time, we must always ask
1 Similarly, it is, in the case of the spending policy so consistently followed by the
government of Louis XV, entirely irrelevant that most modern advocates of "expansion"
would not approve of the objects to which that expenditure was applied. Whatever those
objects, such expenditure ought to have produced prosperity without relapse. If there is
anything in a theory which, in fact, was widely held at that time and has experienced a most
striking revival in our days, it would be but justice to rewrite the history of, say, the minis-
try Aiguillon-Maupeou-Terray in more eulogistic terms than have so far been used.
694
BUSINESS CYCLES
whether our process still persists. The method of deriving an answer is
to locate the postwar period in our cyclical schema, to formulate
the expectations which follow from that, and to see how far they agree
with observed fact. According to that schema the postwar time up to the
world crisis covers parts of the recession and depression phases of our
third Kondratieff which underlie two incomplete Juglars. If fluctuations
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART XL. — German postwar "pulse" (see Appendix, p. 1066).
behaved as they did before the war, those Juglars would be the third and
fourth of that Kondratieff. The third would complete the recession, the
fourth would entirely lie on (but not complete) the depression phase of
the latter. We ought to be able, finally, to discern the Kitchin wave
superimposed on those two. The time series picture of all this must then
link up with the historical facts of the industrial process behind it. Expecta-
tions are perfectly definite and will be formulated as our picture unfolds.
It is recommended, however, that the reader formulate them now and
1919-1929
695
compare them with the Postwar Pulse Charts herewith presented (Charts
XXXIX, XL, and XLI).
•J
^
\
^v-
^/
v^/1
V
\A
HOI
ESA
.E PRIC
s
vxv
S^^s,
/^
"V«—
\
r"
"V
V
sT-
SX^
I
j
i
i
f
1
1
\
IN
'ERE
STI
ATE
i
1
1
\
•1
Iv/
T
V
j
« T"
y
r
"V
»
*
\l
i
1
1
!
i r
1
^
!
\
i
1
Vl
A/
I
i
/
*wj
v/w
^VX
Dt
^U
3OSI
v>w
rs<f
CIR(
:UUA
noN
JvJ
i
i
y^
\S*^
vX
V*^
LA-/
^M
\x^^
^S/
^
^
1
^
If
V
f
^/
ROCUCTION 1
111
vIDE)
V
^
V
V/
\f
i
i
vt
1
\
I
V
1
1
1919 1920 1921 1922 19231924 1925 1926 1927 1928 1929 1930 1931 1932 19331934 1935
CHART XLI. — British postwar "pulse" (see Appendix, p. 1067).
B. Comments on Postwar Patterns. — Now we will drop, if only for a
few paragraphs, the practice, imposed upon us by the nature of our task,
of treating the institutional framework of society, the attitudes of indi-
viduals and groups, and the policies resulting from a given social pattern
as data of our economic process, and changes in these data as external
factors. We will glance at the social process as a whole and in so doing
adopt the convenient, though possibly inadequate, hypothesis of Marx-
ism, according to which social, cultural, political situations and the spirit
in which and the measures by which they are met, derive from the working
696 BUSINESS CYCLES
of the capitalist machine. Our cyclical schema lends itself to this view,
not only because of the length of its longest wave, which brings long-run
social changes within the reach of business-cycle analysis, but also because
it stresses that kind of economic change that is particularly likely to
break up existing patterns and to create new ones, thereby breaking up old
and creating new positions of power, civilizations, valuations, beliefs, and
policies which from this standpoint are, therefore, no longer "external."
The standard illustration is afforded by those innovations which drove the
artisan's shop into modest reservations and, together with the artisan's
shop, also the artisan's world. Gathering up the threads that lie all over
our edifice — see, particularly, Chaps. Ill, VI, and VII — we might thus
try to understand the social configuration of the postwar period from
the economic process we have analyzed.
But we should find the task more difficult than the analogous one
in the cases of the first and second Kondratieffs. There the social
process and its cultural and political complements were not difficult to
interpret in the sense of the working hypothesis which for the moment
we have adopted. All that was not covered by it we could comfortably
stow away as atavisms. This is not so in the case of the Neomercantilist
Kondratieff. If the reader refers to Chap. VII, Sec. E, he will find that
we had to recognize, besides phenomena that indicated consistent develop-
ment of previous tendencies, the presence of other phenomena that did
not seem to fit into the same current but rather to fight against it: they
looked like a revolt against the rational or rationalistic civilization of that
epoch. Of course, it is easy to label them, too, as atavisms. This
sounds convincing in some cases, for instance, in the case of the German
legislation for the protection of the artisan class (Ilandwerkergesetz,
1897). Here we see a dying stratum trying to defend its crumbling
basis by political means. It is not so convincing in others, and any
open mind must admit the possibility that a movement of such breadth
and depth may have been more than an atavism or the last card of a
decaying class. The fact that the writer had no better name to offer for
it than Neomercantilism sufficiently shows that so far he has not succeeded
in interpreting it to his own satisfaction.1 For that term, at best, gets
hold of one of many aspects and is as inadequate as Nationalism or Anti-
1 The atavism theory was expounded by the present writer in his study entitled Zur
Soziologie der Imperialismen, Archiv fiir Socialwissenschaft, 1916. The Marxist theory of
imperialism (Bauer, Hilferding) mentioned in Chap. VII made the attempt to construe
prewar imperialism as an outgrowth of the conditions of trustified capitalism. This
explanation, which conserves unity of principle, has, of course, great attractions for every
mind that has an analytical bent, and could be generalized to include postwar fascism.
It is not possible here to expound the reasons why it is inadequate. A glimpse of a view
that now seems to the writer to be nearer the truth than either the Marxist or his own
theory is embodied in Karl Renner's concept of Social Imperialism (Sozialimperialismus).
1919-1929 697
rationalism would be. Now that tendency or attitude did not perish.
On the contrary, it developed during the postwar period and in develop-
ing revealed itself more fully in the Corporative or Totalitarian or Fascist
State and also became ideologically articulate. However much the war
— and the circumstances of the " Have-not nations" — may have had to
do with concrete forms, mechanisms, timings, and surface events, that
departure from the road that leads from capitalism to orthodox socialism
is not "due" to it, and the general drift of this page might have been the
same had it never occurred.1 The answer to the question how this
development may be expected to affect our cyclical process depends on
the kind of planning that a fascist government undertakes : given sufficient
power and insight in a central authority, innovation may of course be
planned for in such a way as to minimize disturbance.
1. If this component of postwar history can be traced to prewar
sources, everything can. For the only other component — the "socialist"
one — is perfectly en regie from the standpoint of our working hypothesis
and may readily be described in terms of the rationalizing, leveling,
mechanizing, and democratizing effects of capitalist evolution. This is
too obvious to detain us, and only a few points of particular relevance to
our subject need elaboration.
First, the rise of the labor interest to a position of political power and
sometimes of responsibility, which is but the most conspicuous of the
symptoms of a profound change in social structures, is clearly a product
of capitalism in our sense of the term, which created a political world and
political attitudes fundamentally incompatible with itself even where, as
in the United States, the labor interest was (within our period) not
politically dominant. The habit of the old-fashioned liberal — in the
European sense of the word — of blaming "politics" for almost everything
he considers less than satisfactory in the capitalist world is, as far as this
goes, in fact open to the objection that in blaming "politics" he is
blaming a product and an essential element of the system he approves.
Taking the social system of capitalism as a whole, it is meaningless to
1 The question naturally arises whether that turn of phrase is admissible at all. The
answer is that, even from the strictest determinist standpoint, there is meaning to an
analysis — a Gedankenexperiment — of what difference the elimination of an element in the
historical process would make. Another problem may be noticed in passing: it has been
held that great wars are merely incidents of the process of the Long Wave, "regularly"
coming near its upper turning point as a consequence of the expansion of the productive
apparatus which then fights for markets. There is nothing in that "regularity." For it is
obviously absurd to interpret either the Napoleonic or the Franco-German War as explo-
sions of redundant capitalist forces — the French industry was, in Napoleonic times, to a
great extent in the craft stage, and even in 1870 Germany was a primarily agricultural
country. The problem, therefore, arises only in the case of the war of 1914 to 1918. The
writer thinks, though he cannot prove here, that the history of its causation clearly yields
a negative result.
698 BUSINESS CYCLES
say that it — or any element of it, e.g., the gold standard — is checkmated
by " politics." What ought to be said — on this level of analysis — is that
it checkmates itself.
Second, it is worth mentioning that capitalism, also by its own work-
ing, evolves a phenomenon which, or the importance of which, was not
foreseen by Marx : the clerical class. The growth of the laboring stratum
proper, relatively to the increase of gainfully occupied persons, ceased
in the first decade of this century, but the relative growth of the salaried
employee (roughly equal to "white collar") then became spectacular1 —
for obvious reasons of capitalist technique. The interests of this class —
the "logic of its situation" — and its attitudes differing considerably from
the interests and attitudes of "workmen," we have here a factor to the
power of which much of the politics and policies of postwar times may
be traced, particularly in Germany. We must dismiss this most inter-
esting subject with two remarks: that this New Middle Class, as it has
been called, forms in some countries and comes near to forming in
others, together with farmers (peasants) and small businessmen (mainly
retailers), a majority of the population,2 which, though split into widely
differing sections, yet feels and acts uniformly in many cases; and that
in fundamental attitude it is as hostile to the interests of the bigger and
big bourgeoisie as is the working class in the narrower sense of the term,
though also hostile to the interests of the latter. It is in the light of
these facts and not in the light of the simple but entirely unrealistic3
1 Take, for instance, the German case. The figures of 1907 and 1925 are but imper-
fectly comparable. But the number of gainfully employed persons (Erwerbstdtige)
increased by 24 per cent, the number of workmen by 24 per cent (34 per cent in industry,
however; there was a decline of 9 per cent in agriculture) and the number of salaried
employees (Beamte und Angestellte} by 66 per cent, in industry (and Handwerk) alone by
135 per cent. The absolute figure was in 1925 a little over 5.2 millions.
2 Take, again, Germany in 1925: "Independents" (including the members of their
families, whether working in the business or not, and the independents without a profession,
but deducting one million for the "capitalist class") were, roughly, 23 millions; employees
with their families, roughly, 10 millions — #3 millions in all. Laborers (and their families;
we include domestic servants) were not quite 28 millions (total population, roughly, 62
millions).
3 Many sociologists and economists try to keep to that simple schema by means of two
additional propositions: first, that the farmer and small businessman are on their way to
extinction, hence will not, in the long run, help to swell that intermediate stratum; second,
that the white-collar employee is really a proletarian like the manual laborer and ought to
take his true place within the proletarian front. We need not go into the question whether
or not these propositions will turn out true in the long run. For our purpose it is sufficient
to state that they are not true for as much of the postwar period as has so far elapsed,
Ex visu of that span of time, it is, in particular, not true to say that the bulk of white-
collar employees has refrained from joining the proletarian front merely because of the
pressure to which they are exposed. This assertion about pressure seems to rest on a false
generalization from inadequate observation. It is not so easy to exert pressure of the kind
1919-1929 699
contrast between property owners and proletarians that postwar patterns
must be understood.
Third, capitalist evolution not only upsets social structures which pro-
tected the capitalist interests, by progressively eliminating precapitalist
strata from politics and public administration and by creating new posi-
tions of political power, but also undermines the attitudes, motivations,
and beliefs of the capitalist stratum itself. Even if an industrial family
happens to own a given concern, wholly or nearly so, and if its members
actually manage it, they do not under modern conditions look upon it in
the way industrial families used to do in the past. Their attitude is
more distant, less personal, more rationalized. But the leading men of
the giant concerns as a rule fill a specialized function in a spirit which
resembles that of the employee properly so called, and tend to distinguish
between their success and that of the concern, let alone that of the share-
holders. Moreover, the loosening of the family tic — a typical feature of
the culture of capitalism — removes or weakens what, no doubt, was the
center of the motivation of the businessman of old. Finally, the top
group — say, 40,000 men and their families in this country and just
about as many in Germany — absorbs, subconsciously and by an infinite
number of channels, views, habits, valuations — cultural worlds — that
are not its own. "Capitalists" cease to believe in the standards and
moral schemata of their own class. They adopt, or connive at, many
things which their predecessors would have considered not only injurious
to their interests but dishonorable: in surveying modern economic fact,
one cannot but be struck by the discovery of how much of the typical
behavior of the bourgeoisie of the nineteenth century was extra-economi-
cally conditioned. All this, of course, links up, in a way that hardly
requires explanation, with the decrease in the importance of the entre-
preneurial function noticed in Chap. Ill, Sec. C, 5 and Chap. IV, Sec. B.
Fourth, both the rise to power of strata untinged by bourgeois atti-
tudes and the fact that these attitudes lose their hold on the bourgeois
stratum itself, which to an increasing extent allows itself to be educated
by its new masters and the intellectual exponents of these masters — as it
used to allow itself to be educated by its former feudal masters — combine
to produce the anti-saving attitude of our epoch so clearly voiced in its
popular, as well as scientific, literature on economic theory and policy.
Saving with a view to providing revenue for an indefinite family future
was part not only of the economic but also of the moral scheme of life of
the typical bourgeois. The attempt to prove that such thrift is harmful
to the interest of the masses always has been a major element in anti-
capitalist arguments, which without it would be open to a dangerously
contemplated on a body of employees. And nothing can be clearer than the fact that it is
precisely the proletarian class consciousness that is wanting in that stratum.
700 BUSINESS CYCLES
obvious reply. Attempts to prove that it is also harmful to the capitalist
interest itself have never been wanting. But in our time the lesson is
being learned and beginning to motivate public policy. Whatever its
merits or demerits, its success must be understood as part of a general
short-run attitude of modern man toward economic problems and situa-
tions, which follows from the changes in social structure. All this spells a
profound change in the environment within which the capitalist engine
works, and in particular may invalidate in a near future that extrapola-
tion of past performance which, despite the arguments put forth by some
authors in order to establish a tendency toward retardation, it was thought
justifiable to make (Chap. IX, Sec. B; also see infra, Chap. XV, Sec. H).
2. But recognizing thus fully the relation, at least of interdependence,
possibly of causation, which exists between capitalist evolution and its
social and sociopsychological complement, no more debars us from
recognizing the existence of distinguishable spheres of social activity
between which in a given case given effects may be apportioned, than
recognition of universal interdependence of prices debars us from dis-
tinguishing them and from tracing given effects to the behavior of, say,
one of them. Moreover, every such sphere, however much the product
of one comprehensive process, acquires, when once formed, a Jife and
mechanism of its own that enjoys many degrees of freedom. This is
enough to justify our going on to work with the concept of External
Factors. In particular, it is clear that we cannot from a study of eco-
nomic conditions alone determine what will happen in the political
sphere, and if we cannot, there is not much practical value in emphasizing
any quasi-religious belief in such determinateness that we may harbor.
On the contrary, we must deal with all the facts of each sphere as we find
them — which precisely means that they are external factors for one
another. For instance, it is not enough and it is perhaps not even very
enlightening to know that — possibly — England's economic conditions and
interests to some extent explain her general attitude to the United States
during the Civil War. This will not explain why she came within an ace
of, yet refrained from, interfering by the force of arms. Even the
necessity — so repulsive to many students — of taking account of the
personal element cannot be denied. Hence, purely economic diagnosis
of the type for which political action is merely a "disturbing factor" is
not necessarily devoid of meaning but may be instructive to bourgeois and
socialist alike (see ante, Chap. I). We may also note in passing that
there is no justification whatever for the habit some economists have of
skirting uncomfortable issues by pointing to the compelling force of social
and political situations. To any situation the political mechanisms and
their crews can react in many different ways. What is compulsion to
some is not compulsion to others. In any case, it is not his professional
1919-1929 701
equipment that entitles the economist to pronounce about this question.
And the truth or falsity of his economic propositions is as independent of
"political necessity" as the truth or falsity of a doctor's diagnosis is
independent of whether or not the patient is willing or able to act upon it.
In this sense the World War is for us an external factor. From the
facts the above remarks were intended to cover, it seems to follow that
it did not "create" any of the fundamental social features of the postwar
world, although it accentuated some and may have anticipated others.
The physical destruction — including the expenditure of productive energy
on that tremendous "excess in consumption" — and the loss of life in the
most active age groups were made up quickly, the former with a prompt-
ness which in another social atmosphere would have been admired as a
marvel of industrial efficiency. All this reduces for our purpose to two
statements: first, that physical destruction, reinforced by the accumula-
tion of omitted replacements and investments, became the source of a
reconstruction demand, not only in our three or in all belligerent coun-
tries but also in others, which accentuated the prosperities and revivals
that occurred up to, roughly, the middle twenties, although this effect
fully asserted itself only in this country and was in England and Germany
counteracted or delayed by the well-known circumstances to be noticed
presently; second, that the shift from war to peace conditions involved
rearrangements that almost wholly account for the short "jolt" in 1918
and partly for the crisis of 1921, which was, however, in step with the
ordinary run of cycles and was only intensified by this factor (as was the
crisis of 1815).
The moral disorganization brought about by the war and by inflation
and their effects on the cultural inheritance of mankind would, in any social
history of our time, have to occupy a much more prominent place. It
accounts to this day for many of the most striking phenomena in all
sectors of social life, and it was mainly responsible for the lack of stamina
that was displayed by the ruling strata in some countries and that sud-
denly forced issues into practical politics for which evolution was pro-
viding, but had not yet provided, the necessary conditions. One example
will suffice to illustrate this proposition and its bearing on our subject:
no serious socialist would in 1914 have expected — no serious socialist
would have desired — that practical realization of the socialist program
would by 1918 become, in many European countries and in an otherwise
quite immature situation, a political issue, merely because of the moral
breakdown of the ruling strata. When socialist parties were, neverthe-
less, confronted by that fact, they were unwilling or unable to use the
Russian method and to bear down resistance in streams of blood in order
to force into the socialist mould a humanity that would not shape for it
of its own accord. Deadlock ensued in which neither "capitalist" nor
702 BUSINESS CYCLES
"socialist" policy prevailed. As a sort of compromise, the capitalist
engine was allowed to work, but was put under a pressure that prevented
it from working according to design. We shall presently glance at the
case of Germany, which exemplifies that state of things particularly well.
But its fundamental contours also show elsewhere, for instance in
England. It is not generally realized — many people have a strong aver-
sion to realizing1 — how important this is for an explanation of the eco-
nomic history since the war and how many difficulties were created or
increased by the attempt to handle the essentially intracapitalistic
problems of this period by anticapitalistic methods, which yet nowhere,
with the exception of Russia, went far enough to replace the economic
forces they weakened by those of another social system. Short of
socialization, the most obvious measure of the pressure to which the
capitalist organism has been and is being exposed, is the amount and the
progressiveness of taxation.2 Independently of this, the direct effects on
the economic process of income, corporation, and inheritance taxes are
considerable (see below, Sec. C). For both reasons, a prominent place
will be allocated to fiscal policy in the analysis that is to follow. In con-
sequence, other aspects of domestic policies will have to be neglected.
3. Thus, the war undoubtedly precipitated, and gave their particular
forms to, developments which it is reasonable to assume would have come
about without it, but would have come about more slowly and in different
forms. It is in connection with these developments that foreign policies
and the problems of international economic relations in general must be
seen. Their history naturally divides up into three periods : the period
of continuing economic — in some points even extraeconomic — warfare,
from the armistice to 1924 (London conference; Dawes plan); the period
during which the arrangements arrived at were widely believed in and
acted upon, and which lasted from the Dawes plan to roughly the end of
1927; and the period of increasing friction, ending in the midst of the
world crisis in the Hoover moratorium and, later on, in the liquidation,
viafacti, of the reparations and inter- Allied debts. Since the first period
coincides with German postwar inflation, which we exclude, all that need
be said about it is that, while American business situations were for the
1 That aversion is easy to understand. Most people who write and think about postwar
events are in sympathy with socialism of some sort. But most of them are not militant,
so far as the ultimate goal is concerned. It is natural for such essentially "transitional"
mentality to accept for the time being a highly socialized capitalism as a substitute. It
is no less natural that, having accepted it, they try to believe in its possibilities and to
fight against any suggestion that it may be productive of consequences which from their
own standpoint they would have to consider as "undesirable."
2 It also affords, excepting times of national emergency, the most obvious measures of
the political power of the capitalist class. Everything else may be phraseology; the
income and inheritance taxes are facts.
1919-1929 703
time being little and, if anything, favorably affected by the course of
international events, England suffered greatly: her exports to Germany,
for instance, were even after 1924 smaller by about one-third than they
had been in 1913 (measured in the 1925 units of the Statistische Reichs-
amt), but before that year her troubled relations to her "best customer*'
certainly helped to account for her share in the depression of 1921 and for
he unsatisfactory recovery that followed.
As regards the second period, our statement that the arrangements
arrived at were acted upon requires qualification. The world never
squarely faced their fundamental consequences, the unavoidable Ger-
man exports in particular, but poured American and other credits on all
the unsolved problems and plastered gold currencies — most curious
atavisms though they were, in a world otherwise resolved not to play the
capitalist game — on essentially untenable situations. Since it is impossi-
ble to enter upon these matters and since the salient facts may be assumed
to be familiar, we can for our purpose compress the comments that would
be necessary into the statement that this "export of American capital"
did not act as we would expect capital export to act under more normal
conditions, but that for the time being it largely counteracted the dis-
turbance that would otherwise have ensued from international political
payments. Moreover, while it thus relieved stringencies in many coun-
tries, it did not create corresponding ones in the lending country, because
America still remained a creditor on current account, even apart from
the short balances which for familiar reasons continued to flow toward her.
This explains the astonishing fact that those "political transfers" will
not play any great role in our analysis of the processes of that period
except in the German case, and that even in the latter case it is transfers
to Germany, rather than transfers from Germany, that shaped situations.
This even applies to the third period, almost until the outbreak of the
world crisis when, though partly for political reasons, short balances took
fright. But the fundamental facts of the situation, accentuated as well
as expressed by protectionist tendencies and the passing of the Locarno
atmosphere, had begun to assert themselves before that.
It should be observed, however, that if those temporary solutions of
the international financial problems created by the war proved inade-
quate, and if their economic consequences, including their — secondary —
share in the causation of the world crisis, turned out as they did, this was
only because of the political environment within which they had to work,
i.e., that our assertion of their inadequacy must be understood to be
relative to the social situation previously glanced at. They were bankers'
solutions, which the nations concerned were unwilling to accept and
which they defeated by refusing to allow the mechanisms to work on
the functioning of which the authors of those solutions relied. Looked
704 BUSINESS CYCLES
at as business propositions, they would not, in a peaceful world accepting
bourgeois standards, have been obviously absurd, and it would not in
such a world have been unreasonable to expect that they would work out
in the end. The well-meant proposals of all those international con-
ferences that met to discuss the gradual removal of trade barriers look to
us like strange anachronisms and certainly were as futile as proposals for
disarmament must be in a world in which every nation that counts is bent
on armament. But they were perfectly sound economics. JEven the
gold currencies were such failures only because trade barriers, fiscal
policies, social and military expenditure, and insistence on higher money
wages did not allow them to function and because in that hostile environ-
ment short capital was rushing about like a hunted hare. Given all
these facts, it was and is indeed little short of ridiculous to trust to the
remedial forces of laixser faire. But since they do not, any more than
the war itself and Versailles, uniquely follow from the logic of our evolu-
tionary process, it is not to the interest of clear thinking to speak of any
inherent tendency of capitalism to run into such deadlocks.1
4. We will mention briefly a few examples of various other types of
war effects. In some countries, such as New Zealand, war demand had
induced expansions which were clearly untenable under normal conditions.
Even disregarding the speculative excesses that actually occurred, we
have no difficulty in understanding that liquidation was unavoidable.
In New Zealand it began in 1921 and continued, witness the abnormal
number of bankruptices, almost without interruption right into the
world crisis. Protection, borrowing, and prosperities in many parts of the
world mitigated the process, which up to the crisis spelled additional busi-
ness for England — the total imports of New Zealand more than doubled
from 1919 to 1920 and then remained on a lower, but still very high level
— and in the crisis additional embarrassment through frozen credits.
1 The reader need not fear that the above argument proceeds from any intention to
"whitewash" capitalism. For, as pointed out before, it leaves him free to condemn as
much as he pleases the social system as a whole which produces such deadlocks. But it
serves no purpose to blur the diagnosis of that situation by faulty economic argument.
This has been done not only by those writers who are prisoners of a rigid formula, but also
by economists who care more about conveying their practical advice than about analysis.
The discussion on the reparation problem may serve as an example. Instead of stressing
the simple and trivial truth that the Dawes tribute was morally inacceptable to Germany,
and economically inacceptable to the recipient countries — which would never have sub-
mitted to having their industries crippled by the German exports that would have been the
necessary consequence of actual payment — many, especially English, writers tried to prove
the economic impossibility of those exports on grounds of elasticity of demand, while it is
perfectly clear that, looked at as a business proposition, the Dawes tribute would have been
nothing else but a "commission" paid by Germany for the industrial conquest of the
better half of the world. If the world was unable to act upon this possibility, it may have
been sound advice to cancel both reparations and inter-Allied debts, but this does not
save the reasons by which that advice was supported.
1919-1929 705
The example is really typical of many more important cases, all of which
contributed their special crises to the world's economic situation after
1929. It is worth mentioning that in the maladjustments revealed by the
cessation of war conditions, the New Zealand wage level cannot have
counted for much. If the official index may be trusted, real wages moved
under the 1914 level until 1928. This also applies to other cases. The
widespread opinion, for instance, that Australian wages were a major
cause of subsequent troubles is hardly borne out by observable facts.
In other countries the war quickened industrial expansion, which
it is reasonable to assume would have ensued anyhow in the course of
time, and thus contributed to that decline of the position of Northwestern
Europe in the world's economy, which in itself is not a war effect. The
rise of a native capitalism in the tropics is a case in point. The war
profits of India, for instance, gave an impulse to native industry which,
since it directly affected Lancashire, may be listed among the factors
which shaped England's economic history during the period. But the
effects of all the other cases of this type that might be mentioned — Japan's
great stride in industrialization, for instance— either dispersed themselves
over too wide a surface to signify for our purpose or were, for the time
being, too much compensated for by the demand for foreign goods —
investment goods; a characteristic change occurred, as compared with
prewar times, in the composition of European exports — incident to the
early stages of the process.
Still another type of consequences may be instanced by the breakdown
of Russia and the subsequent developments under the Bolshevist rule.
It is obvious that these events, had they occurred in an otherwise "nor-
mal" world, would have greatly affected conditions in general and the
cyclical process in particular in several other countries, particularly in
Germany and England. As it was, however, the effect was lost in a
mass of much greater dislocations of industry and trade, and no distinct
process of adaptation to the new state of things in Russia was required.
Business all over the world had, of course, partly to go without the
stimulus that Russian reconstruction and expansion would have offered
under other circumstances: if we extrapolate her industrial development
in the 16 years preceding the war, we have no difficulty in forming an
idea about the quantitative importance of this lucrum cessans. More-
over, without the loss of French investments in Russia, understandings
about reparations and inter-Allied debts would have been much easier.
But it should not surprise us that we find little of positive effects. Much
the same argument applies to the case of China and to others.
5. Since the role which postwar protectionism played in the develop-
ments of the twenties and in the causation of the world crisis has been so
completely overlooked by some students of the business cycle and so
706 BUSINESS CYCLES
obviously exaggerated by others, it will be convenient to restate explicitly
the view adopted for the purposes of our analysis — which, it should be
borne in mind, excludes the wider aspects of the matter, such as the rela-
tion of protectionist policy to human welfare and to peace. In the first
years after the war, duties, prohibitions, quotas, and other weapons from
the arsenal of protectionism were, of course, elements of the general
scheme of continuing economic war. But they were also something else.
Adaptation of industry and trade to the new conditions above referred to,
permanent and transient, was at best a difficult task, involving* in many
cases abrupt dislocations. This becomes obvious if we glance at the
postwar figures both of commodities produced and of commodities
internationally exchanged. In some of these cases, protective tariffs or
even prohibitions were, if not the only means, yet the most obvious
means of averting sectional catastrophes, from which cumulative proc-
esses (spirals) might easily have ensued. Unequal depreciation of cur-
rencies, of course, greatly added to this class of difficulties. Many
measures, such as the McKenna duties and even the Fordney-McCumber
Act, must, in part at least, be interpreted in this light and on balance
probably mitigated many more difficulties than they created. At the
other end of our period, immediately before and during the world crisis,
a similar argument applies, especially after the newly established cur-
rencies had begun to give way, although with lesser force and although
a panicky policy of protection and " incapsulation " then went to obvi-
ously irrational lengths.
In the years from, roughly, 1924 to, roughly, 1928 some steps toward
freer trade were actually made, some countries removing certain barriers
and the tariffs of others being automatically lowered, in cases of specific
duties, by depreciation not always compensated for by a gold clause.
However, it is understandable that not more was accomplished: disloca-
tions and untenable war growths continued to exist; unequal depreciation
of currencies was replaced by unequal stabilization, which in some cases
overvalued and in others undervalued the legal tender unit; political
payments, especially but not only in the case of Germany, provided a
motive for aiming at an active balance of trade entirely justifiable
even from a free-trade standpoint. There was, indeed, "nationalism."
Its outstanding manifestations are to be found in the policies of the
majority of the newly created small states that tried to foster industrial
development at any price. But that great movement of which we
primarily think when speaking of modern nationalism, and which has
been recognized above as allied to one of the two great components of
the social atmosphere of today, has really little to do with the commercial
policy of the great nations — and the small ones of old standing — during
those years. That policy was dominated by current vicissitudes partic-
1919-1929 707
ularly in England, America and — until the National Socialist party rose
to power — in Germany. England substantially adhered to the principle
of Free Trade until 1932. The edge of German protectionism was in her
agrarian situation. And again it is reasonable to think that on balance
and in spite of many ill-conceived individual measures, European pro-
tectionism during those years hardly accentuated prosperities and
depressions, or created on balance additional maladjustments of major
importance, whatever we might have to say about it from other stand-
points. Moreover, it must not be forgotten that the volume of inter-
national trade could in any case not have been expected to develop
along prewar trends, and that a certain amount of "autarky" was the
unavoidable consequence of technological progress in many branches and
countries (see Chap. XIII, Sec. C). Electricity and chemistry alone
would have been sufficient to produce a tendency in that direction, so
that the actual course of events cannot be simply attributed to protec-
tionist barriers.
This country's famous "refusal to accept its creditor position'* still
remains. In that phrase there is, no doubt, some element of truth.
But since, as pointed out above, American capital export to Europe — it
reached about 5 billion dollars by the end of our period— more than
sufficed to service her claims, the consequences for the time being were
merely to quicken reconstruction in Europe. It was not this mechanism
that produced the world crisis, but the world crisis that caused its break-
down. Then, of course, the situation thus created became a major
factor in the ensuing depression. But even then it is not easy to see how,
had a reduction of import duties been passed instead of the Hawley-
Smoot Act, this could have improved short -run conditions in Europe
without aggravating them in America. Whatever the merits of free-
trade sermons, they can only apply either to the course of action that
might have been followed had the crisis not occurred, or else to the course
of action that might have been followed after it had passed. With this
we are not concerned. For our purpose it suffices to conclude that pro-
tectionism as such played but a minor role in the cyclical process of the
postwar epoch, and to cast a glance at this country's international balance
of 1928, since that is the last complete year of the "prosperity plateau."
Commodity exports to Europe were then $2,342,000,000; imports of
commodities and services from Europe plus remittances of immigrants
and tourists' expenditure, very roughly — estimates of tourists' expendi-
tures are particularly unreliable — 2 billions. The resulting net credit
of between 300 and 400 millions has to be increased by payments received
on war debt account — 200 millions — and net receipts from interest and
dividends(P). This makes about 600 millions, which must have been
mainly "paid" from additional credits, since the total of monetary gold
708 BUSINESS CYCLES
in the United States had fallen both in 1927 and 1928. It is true that
the immigration of short funds, then approaching their 1929 peak, com-
plicated the situation. But, with due respect for the excellent motives
behind many of the exaggerations of which economists of all countries
were guilty in the matter of American policy — Europe's willingness to
lecture was more obvious than her willingness to pay — it is presumably
safe to say that a sum of that order of magnitude could not, without a
crisis, itself caused by other factors, have created an unmanageable prob-
lem. In the ordinary course of things, adjustments of the commodity
balance amounting to a sum of the order of 300 millions would have been
possible, even in a protectionist world: exports alone could have been
gradually reduced to that extent without serious repercussions on the
American situation,1 while reinvestment could have absorbed the rest.
Again, it was the crisis that prevented such adjustments, and suddenly
made an insoluble problem of what otherwise was not only not beyond,
but on the way to, a solution which, though in the future it might have
entailed, did not then presuppose free trade.
C. Further Comments on Postwar Conditions in Our Three Coun-
tries.—
1. More nearly than any other country, the United States displayed,
and substantially retained until the world crisis, a frame of mind appro-
priate to the task of running the capitalist machine, even to the extent
of reducing what was almost universally disapproved of as an "un-Ameri-
can radicalism*' to still smaller importance than it had had before the
1 This reduction would have been brought about automatically by reduced capital
exports if the meaningless policy of fostering commodity exports had been discontinued.
Some planning would, however, have been necessary in order to safeguard those sectors on
which the reduction of commodity exports would primarily have impinged. For it must
be remembered that while American exports did not greatly signify as compared with the
total of the nation's business, they did signify for agriculture, textiles, metals, machinery,
and vehicles. It should also be remembered that we are discussing relations with Europe
only. The relations of this country with other parts of the world raise different problems,
which cannot be discussed here. It is obvious, however, that as far as the period under
discussion is concerned the analogy with the English case in the second half of the nine-
teenth century is misleading, because England's established creditor position then bore a
different relation to the additions that were made to it in any given year. A protectionist
policy would, therefore, have created many more difficulties in her case than it can have
created in the case of the United States : while a creditor position is rapidly being built up,
the lending country tends to be in the position of debtor. Needless to say, nothing of
this is, or intended to be, an argument for protection in general. But one more error
about this country's creditor position should be mentioned, viz.t the error that it was due to
the war and would not have developed without it. This is true only of the 11 billions of
"political" claims and beyond them for the speed of the process. But fundamentally
the United States were well on the way out of a debtor nation's and toward a creditor
nation's position in the last prewar decade, and it is reasonable to assume that they would
in any case have reached the latter by 1920.
1919-1929 709
war. Such deviations as occurred from those principles of action that
are associated with the logic of the capitalist process were due, rather than
to the intrusion of ideas hostile to that logic, to the failure to adapt old
ideas to the new situation, as exemplified by so much as there was in the
"refusal to accept the implications of this country's new creditor posi-
tion" just discussed.
But apart from this and possibly monetary policies (see infra, Sec. F)
and the rate of municipal expenditure, there were no lesions inflicted on
the system by action from the political sphere. On the contrary, while
the nation was bending its energies to the type of tasks characteristic of a
Kondratieff downgrade, the federal government was pursuing a fiscal
policy eminently "sound" in the old sense. It reduced taxation, going a
considerable way beyond merely eliminating the excess profits tax; it
reduced the federal debt and even set about to effect some retrenchments
— Gladstone himself could perhaps have acted more brilliantly, but
hardly more soberly. Up to an income of $100,000, the income tax
was far below the European level. Federal expenditure, which in 1912
-1913 had been 724.5 millions, moved, it is true, on a level of about
3.7 billions from 1925-1926 to 1929-19301 (including debt redemptions
out of current revenue to the average amount of over % billion a year2).
But under general conditions so exceptionally favorable this was not a
very serious matter. Local Total Gross Expenditure increased from
4,593 millions in 1923 to 6,720 millions in 1929; State Total Gross
Expenditure, from 1,208 to 1,943 millions. But both states and local
authorities raised, partly from a lack of constitutional powers, partly
from choice, the money they spent in ways which (whatever may be
1 Federal Total Ordinary Receipts (as per Treasury Statements) were: 1920, 6,694.6
millions; 1921, 5,624.9; 1922, 4,109.1; 1923, 4,007.1; 1924, 4,012.0; 1925, 3,780.1; 1926,
3,962.8; 1927, 4,129.4; 1928, 4,042.3; 1929, 4,033.3; of which Tax Collections (National
Industrial Conference Board, Cost of Government, 1923-1934, p. 18) were: 1922, 3,487.0;
1923, 3,032.0; 1924, 3,193.0; 1925, 2,966.0; 1926, 3,207.0; 1927, 3,337.0; 1928, 3,194.0; 1929,
3,328.0. Total Gross Expenditure was (Treasury Statements) 6,482.1 millions in 1920
and steadily fell to, roughly, 3.5 billions during 1924-1927, after which it rose again (1929,
3,884.5; it was 5,153.6 in 1932).
2 Those redemptions were about compensated by the increase in state and municipal
debt. Total public debt decreased (slightly) only in 1923 and for the rest of the period
continued to increase, although total public issues outstanding, minus holdings of the
United States government and government trust funds, slightly declined, moving on a
level of about 30 billions. Therefore, we need not bother about the question how net
redemption would have affected money-market situations and, through them, speculation
and business. That would have depended on the sources of the funds by which redemption
was effected and on the behavior of the households, firms, and banks that held the bonds
redeemed. By drawing up a schema of possible combinations, the reader should have no
difficulty in developing a complete theory of the effects of debt redemption and in realizing
that both "deflationary" and "inflationary" consequences may ensue from it.
710 BUSINESS CYCLES
thought about them from other standpoints) did not substantially
injure the economic machine.
Moreover, the government promptly abolished most of the wartime
controls, regulations, and organizations; refrained from measures involv-
ing questions of social and economic structure at home; and successfully
kept out of entanglements abroad, thereby creating the atmosphere con-
genial to private business1 and reducing the importance to the American
citizen of the struggles, sufferings, and upheavals in other p$rts of the
world to the order of importance of a football match. Economists who
are passionately determined not to admit that policies answering to their
social and moral vision, particularly fiscal policies of anticapitalist
tendency, can possibly interfere with the working of the economic system,2
will no doubt hold that there was mere chance coincidence between that
sociopolitical pattern and the economic results achieved in this country
during the twenties, and between the different setup and the different
results in England or Germany. In the fulfillment of our humble task
of interpreting a given course of historical events and the behavior of
given time series, we cannot, however, neglect the possible inference to
the contrary. We speak of possible inference only, because in this point
our argument transcends exact proof, as any argument about organic
processes occasionally must, and because so many imponderable elements
enter which must be a matter of personal judgment and (historical and
personal) experience.
2. But the main points at issue with reference to effects of taxation
as such — i.e., as distinguished from those effects which a system of
taxation may have if it is or is felt to be an element of a general atmosphere
of hostility to capitalist success3 — may conveniently be mentioned here
1 It has been pointed out to the writer that the above argument reads like an advocacy
of partisan principles. He does not see that. Nor does he see how he could guard against
this danger, short of inserting explanatory parentheses to an impossible extent. If the
reader misunderstands that passage, he will misunderstand so much in this book that
one more misunderstanding hardly matters. Let it be reemphasized, however, that the
above (which, besides, contains only statements of obvious fact to which nothing will be
added except an inference about a causal connection) neither is intended to convey, nor is
as a matter of fact sufficient for, an appraisal either of capitalism or of the particular
line of action taken by certain political groups within a given historic situation.
2 They could, however, as we shall not tire of pointing out again and again, admit it
without prejudice to their standpoint, which obviously rests on extraeconomic valuations.
What is an injury to an organism from the standpoint of a doctor, may be most "desirable"
from plenty of other standpoints.
8 To this aspect we shall have to return in the last chapter. Also, the above remarks
cover only part of what for our purpose it is necessary to say on the subject. The rest has
been divided up and inserted in various places in such a manner as to minimize expenditure
of space. It includes some remarks on the problem how the use made of the sums raised
by the receiving authority influences the effects of the "burden," on the view that taxation
for certain purposes is "mere transference," and on other topics.
1919-1929 711
once for all. Since we cannot fully go into the matter, we will avail
ourselves of the fact that there is (comparative) agreement about the
effects of indirect taxes, such as specific taxes on the quantity produced
or sold of a commodity. This agreement, such as it is, we owe to a fairly
well elaborated theory which, though antiquated, is still widely accepted
by economists and has recently been somewhat improved by borrowings
from the theories of imperfect competition, of expectation, and so on.
Its assumptions, however, limit its results to the case of small taxes and /or
of individual commodities of small importance. The technical reason
for this has an important counterpart in real life: wherever taxes are so
small as to be amenable to analytical treatment by the calculus, they are
also too small to affect the fundamental contours of economic behavior as
reflected in the budgets of firms and households and, hence, to interfere
significantly with economic processes in general and the cyclical process
of evolution and its permanent results in particular. This proposition
may be generalized to cover any small tax, no matter whether sectional,
such as a tax on beer or on house room, or general, such as a turnover or an
income tax, and extended in most cases to any tax that is small in a prac-
tical— though loose — sense and not only in the sense of the calculus.
Most taxes — not strictly all — which are not small in that wider sense, on
the one hand cannot be handled by that method — further repercussions,
more fundamental changes in the economic system, reactions from and
through the sphere of money and credit must be then taken into account
— and on the other hand, do interfere with the results of business proc-
esses, for example, with the steady rise in the standard of living of the
masses as far as it is due to the working of the capitalist machine.1
This, however, marks the point at which disagreement begins. The
fiscal problem of our time does not primarily consist in the amount of
revenue required by the modern state, but in the fact that, owing to the
moral valuations prevailing, that amount must also be raised by heavy
taxes2 and, moreover, by heavy taxes framed not only without a view to
1 Within limits, other methods of attaining a maximum standard for the masses may be
effective without interfering with the contribution made automatically by the capitalist
machine — the machine can be made to yield more than it would of itself without materially
losing in efficiency. Even if it does lose in efficiency, the advantage to the labor interest
may more than balance the loss it suffers, especially in the short run, but also in the long
run. All that the proposition of our text amounts to is that any such advantage is gross
only and must be subjected to scrutiny as to concomitant losses to be deducted. It
stands to reason that the advantages will, as a rule, be visible and immediate and the
losses more difficult to see and more remote in time.
2 "Big" revenue need not imply "heavy" taxes. A transaction tax, for instance,
which in this country has been actually proposed in the McGroarty bill, would go far
toward raising adequate revenue without significantly altering the conditions of economic
progress. According to Mr. Golden weiser, total monetary transactions amounted in 1929
to about 1,200 billions (debits to individual accounts in 141 cities, plus an estimated addition
712 BUSINESS CYCLES
minimum disturbance but regardless of disturbance, in some cases even
with a view to maximizing it. And the disagreement that is relevant to
our purpose concerns either the reality of the effects alluded to in the last
sentence of the preceding paragraph or their importance for, let us say
a potiorij the development of total output. We will confine ourselves to
the case which is most important in this connection and consider a high
and highly progressive income tax — by which, to fix ideas, we will mean
an income tax which, for a significant number of taxpayers in the higher
and highest brackets, surpasses 25 per cent — that so defines income as to
include savings and is reinforced by a significant corporation and a high
or highly progressive inheritance tax.
First, there are what we may term mechanical effects, of which the
most important is the effect on the sum total of private savings and
accumulations. Taxes such as those we have in mind may enforce dis-
saving and even divestment, but will in general be partly paid from
revenue that, in turn, would otherwise be partly saved. An obvious
argument from general principles yields the result that, as a rule, this
again will be partly made up for by additional saving by the same people
or by those who are the ultimate recipients of the sums levied. But as
far as the writer knows, nobody has so far doubted that the net effect of
high taxes on the higher incomes will be a decrease of the national total
of savings as compared with what it otherwise would be. As far as this
goes, therefore, our opinion on how such taxes will affect "progress"
and "industrial efficiency" depends on where we stand in the controversy
about the importance and the modus operandi of private saving, which
have been fully discussed before.1
for checks drawn on banks outside those cities and payments in cash). % per cent of this
sum or, if certain classes of payments were to be excluded, 1 per cent of half of it would be a
small tax in the sense that it would nowhere exert perceptible pressure. But it would
yield — in a year like 1929 — 6 billions.
Big taxes interfere with the economic process to greatly varying degrees. Big taxes
on individual spots, such as a heavy tax on alcoholic drinks, does interfere significantly
with the particular sector on which it impinges but not with the system as a whole. Among
"general" big taxes, a tax on the return to or value of natural agents would be a good
instance of a limiting case in which no long-run effects are to be expected if it were not for
the facts that improvements are not easy to distinguish from "pure rent" and that the
latter, as we have seen, for example, in the case of the apartment-house industry of Berlin,
steps into the place of and acts as a substitute for profit. But all that matters here is the
technical possibility of raising "big" revenue with small disturbance if that were intended —
and small revenue with big disturbance.
1 But it should be observed that many arguments turn, not on saving in our sense,
but on underspending. Taxes on idle funds may have some stimulating short-run effects
if conceived as temporary measures. This point, too, has been fully discussed in Chap.
XI. One more aspect of saving and accumulation, which is important for an appraisal of
short-run effects, will be added in the next chapter,
1919-1929 713
Second, there are the nonmechanical effects, i.e., the effects through
motives and attitudes. It should be obvious that any tax on net earnings
will tend to shift the balance of choice between "to do or not to do" a given
thing. If a prospective net gain of a million is just sufficient to over-
balance risks and other disutilities, then that prospective million minus
a tax will not be so, and this is as true of a single transaction as it is of
series of transactions and of the expansion of an old or the foundation
of a new firm. It should be equally obvious that business management
and enterprise, being undertaken within an institutional framework of
aims, ambitions, and social values fashioned to its logic, will for its
maintenance depend, at least in the long run, on the actual delivery, in
case of success, of the prizes which that scheme of life holds out, and that,
therefore, taxes beyond a percentage that greatly varies as to time and
place1 must blunt the profit motive and, especially, the motive typical of
both feudal and bourgeois society, that of founding a family position.
As to the profit motive in general, it must be borne in mind that a policy
of taxing away gains evidently above what would be necessary to call
forth the efforts of their individual recipients and of taxing but moderately
what the community considers "adequate" returns, if it is not to affect
the total amount of effort, would really have to be accompanied by an
increase in the sum total of managerial and entrepreneurial income,
because the presence of conspicuously high and even fantastic individual
prizes is, as everyone knows, much more stimulating than the same sum
would be if more equally distributed among businessmen. As to that
special form of the profit motive which is embodied in the term family
position, and is largely eliminated by inheritance taxes of the modern
type, it is as reasonable to hope that high inheritance taxes, being taxes
on "static" wealth, will not affect industrial "progress," i.e., the creation
of new wealth, as it would be to hope that a prohibitive railroad fare will
not affect traffic if passengers be allowed to board the trains free of charge
and the fare be collected from them after they have taken their seats.
The reader has a right to object to the triviality of these consider-
ations. Yet it is a fact that the reality of effects through motives and
1 Moderate taxation, i.e., taxation which, while making it more difficult, yet does not
make it too hard to attain a given economic position, may even act as a stimulus. But
however difficult it may be to determine the interval for which that is so, it is perfectly
clear that since the war taxation in the higher brackets goes much beyond it. — High taxa-
tion, for example in a national emergency, as long as it is considered to be temporary, may
have no effect on motive or even an effect that is stimulating. What taxation is "high" and
what "moderate" also depends on the prevailing margins of profits. American taxation
even from 1924 to 1031 might have been high in our sense but for the ease with which the
businessman rode to success. Finally, much depends on the reaction of the monetary
system, for example, on whether or not taxpayers are willing and able to borrow the
amounts they have to pay.
714 BUSINESS CYCLES
attitudes is still more frequently denied than the bearing of accumulation
on industrial efficiency.1 But there is at least an argumentum ad hominem
to be addressed to any economist who uses the profit motive in his analysis
at all.2 For net revenue would have to be a matter of complete indiffer-
ence if rates such as we now discuss had no effect on the working of the
capitalist machine. If, moreover, that economist teaches, as he is likely
to do, the omnipotence of comparatively small changes in the rate of
interest and the effectiveness in stimulating industry of protection, sub-
sidies, and * deflationary" increases of prices, and still denies that high
taxation has any effect whatever on the scale of output or "progress/*
he comes pretty near to contradicting himself, unless he holds that that
effectiveness completely vanishes at the watershed between profit and
loss.
3. The case of Germany is less easy to describe than that of the
United States. Refraining from a repetition of facts that are but too
well known (although some are perhaps already forgotten) and taking
our stand on the year 1925, we may thus characterize the situation.
Menace of social breakdown had been warded off by the determined
action of the Social Democratic party in 1918 and 1919, which thereupon
found itself in the impossible position of being (whether in office or not)
the dominating political power in a capitalist society which it had saved
when all that society's political organs were utterly shattered, but which
it was by virtue of its principles unable to run according to capitalist
logic.3 The result was that economic policy could not be rationalized
1 Many economists would, however, argue that the effect of taxation on the motives
or the behavior of businessmen, whatever it may be, can be neglected in an analysis of
variation in output and employment, because these motives do not matter anyhow, the
businessman's hunt for profit having nothing to do with either output or employment
which result from the economic process except that it sometimes interferes with them. A
view of this type is largely a vision or impression — easily contracted in this newspaper
world of ours, in which words are in the saddle — about which it is difficult to argue except
by discussing the individual facts that in each case would be produced in response to a
request for substantiation. In a general way, however, our analysis as a whole supplies
the elements for criticism of this vision as well as the elements of its sociology.
2 Marshall's great shade cannot be conjured up to testify against that. For though
he seems to have held the opinion that a liberal dose of direct taxation would not reduce
industrial efficiency, the rates he visualized were, except for temporary emergencies, so
low as to make his argument quite compatible with ours. Nor can it be urged that we are
using a rational schema of behavior that does not take account of fixed habits. We do take
account of them. But this particular application of economic rationality is surely not
unrealistic or farfetched. And those habits are likely to give way under the influence of
unfavorable experience and pessimistic anticipation, as do even the habits of a puppy,
which does not indefinitely jump for a sausage that is pulled out of his reach each time.
Were it not for the reluctance to admit unpalatable facts, there would be little disagreement
about this.
8 Socialists went to lengths, however, which amounted to an opus super erogationis.
The Sozialisierungsgesetz (1019) meant the shelving of socialization. But more was to
1919-1929 715
either in the socialist or in the capitalist sense: a deadening laborism
threatened everybody and satisfied nobody.1 Menace of national
destruction had been warded off by the guarantees against invasion
included in the Dawes plan. Menace of economic deadlock had been
warded off by those credits which the world pumped into Germany, obvi-
ously believing that credits and democracy were all that a nation could
possibly need. Menace, finally, of irreparable moral disorganization
through inflation had been warded off by an energetic, if clumsy, bal-
ancing of the budget, consequent upon the construction of a rather rigid
gold standard, both the Dawes loan and the successful bluff of the
Rentenmark being minor, though still important, safeguards or pieces of
technique.
The complete loss of foreign investments and of industrial and com-
mercial positions abroad would in any case have been sufficient to
unbalance an economy so largely based on them as the German was.
Moreover, Germany's industrial organism emerged from the war with an
antiquated productive apparatus, which then suffered further injury
and, in part, further distortion because of that kind of irrational invest-
ment the sole motive of which was to take shelter from inflation — a good
illustration of the popular belief that if only there be expenditure it does
not matter what direction it takes. A crisis of adjustment both of
structures and of values was bound to set in to clear the ground as soon
as the support of progressive inflation was withdrawn, but much mal-
adjustment then in existence was, as it always must be, very slow in
dying. To the end of the period some of the growths of inflation con-
tinued their sickly life, constituting weak points, sources of chronic
follow. The minister Hilferding, much too good an economist not to see what was wrong
and much too good a Marxist not to realize that there are situations in which anticapitalist
policy is in the end antisocialist, actually went so far as to attempt a very "capitalistic"
fiscal reform. In November 1927 the leading socialist newspaper advocated reduction
of the income tax (in all brackets). Nothing came of it. Difficulties for the socialists
were greatly increased by the fact that their political allies, the centrist party, though less
radical on principle, were much less amenable to economic reasoning than were the socialists
themselves.
1 The picture cannot be developed here, but two features must be mentioned in passing.
First, the trade unions secured two points of their program — the eight-hour day and unem-
ployment insurance. These measures were, in fact, overdue. But beyond that a huge
legislative and administrative apparatus was built up by the competitive exertions of the
two ruling parties for the purpose of serving the immediate interests of labor (the Arbeits-
rechi), which, whatever its merits, is relevant to our subject by virtue of the economic waste
and friction it created, and particularly by virtue of the efficiency with which it "skimmed,"
as the Minister Braun — a centrist — put it, the results of every upturn in business for the
immediate benefit of labor, i.e., applied potential capital to consumptive purposes. Second,
the socialists had to accept a no less wasteful agrarian policy, including subsidies to eastern
agriculture, which weighed heavily on the economic process, yet entirely failed to reconcile.
716 BUSINESS CYCLES
trouble, and one of the difficulties in Germany's economic relations with
other countries.
The weakness of her international position made her industries
particularly anxious to enter into agreements with foreign competitors in
order to eliminate their pressure for hostile measures. Some temporary
success was achieved1 but, as far as it was, it carried all the disadvantages
incident to "stabilization" in creating rigidities and cramped situations.
From 1925 to 1928, however, the surface of things was smoothed by
those foreign credits, by comparatively moderate impediments to trade,
and by comparative order in monetary systems all over the world, as
pointed out above. Even then the risk of industrial investment remained
in Germany much greater than it was in England or in the United States,
and much waste of effort and resources ensued from the necessity of
shifting locations of industries according to political fears, and of remodel-
ing equipment in response to foreign measures — an element never fully
appreciated in diagnoses of the German situation. But when the flow of
foreign capital ebbed and in 1929 finally ceased, and Germany's balance
of commodity trade promptly and easily (without any "stickiness")
swung in her favor in consequence2 — a most striking verification of classic
theory of foreign trade, as was also a temporary tendency in this direction
in 1926 — that smooth surface went to pieces. German reparation
exports had suddenly become a reality. They were one of the influences
that shaped the international commercial situation on the threshold of
the world crisis, helped to induce the vicious circle of mutual restrictions
(quota and so on), and effectively brought out the absurdity of insisting
on payments which the creditor countries at the same time stoutly refused
to accept.
As long as they continued to flow, however, foreign credits smoothed
not only the difficulties incident to transfers on reparation account but
other difficulties as well. In order to understand these it is necessary
to recall what has been said about the social atmosphere and the political
1 By the end of 1926 a considerable list of international cartels was in force, such as the
potash convention of Lugano, the European union of bottle manufacturers, other European
unions of enamel works and wood-screw producers, the international bulb syndicate, the
international rail cartel, the cartel of German and Czechoslovakian iron -pipe producers,
the German- Belgian wire cartel. International organizations for iron, copper, benzol,
and other articles were at the same time under consideration, and some of them actually
matured. After 1927, when French interests became increasingly restive under the
impact of German exports, machinery was provided for producers of both nations to meet
and to find themselves a modus vivendi (quota) by mutual agreement, to be ratified by the
respective governments. For some time this, too, seemed to work with comparative
success.
2 Export surplus: minus 1,725 million marks in 1928; plus 36 in 1929; 1,642 in 1930;
2,872 in 1931. After that it began to fall owing to the universal "mcapsulation" incident
to the world crisis.
1919-1929 717
structure of that epoch, and to look at the fiscal policy that resulted
therefrom. The first federal budget after inflation, though it intensified
the "crisis of stabilization" by cutting deeply into the working capital
of industry, was a signal success, yielding as it did a surplus of about one
billion marks. In 1925 the total expenditure of the federal government,
the states, and the municipalities, including social insurance and repara-
tion payments, amounted to 17.3 billion marks or (as officially estimated)
about 31.9 per cent of the national income, as compared to 8.4 billions
or 18.9 per cent (postwar territory) in 1913. l But Germany's new political
structure was unable to withstand popular demands and to plan rationally
for the future. As soon as the immediate emergency was overcome, and
the unwieldy mass of fiscal legislation systematized and somewhat
adjusted by the "reform" of 1925, that fact asserted itself exactly as it
did in France after 1928. Notwithstanding the friendly help which the
Reparation Agent extended to the government by means of unfriendly
notes and reports, expenditure increased by leaps and bounds. It
reached the figure of 23.3 billions in 1927 and from that year on entailed a
deficit, from 1929 even an embarrassing shortage of cash, although
public revenue steadily increased through 1929, in spite of drastic reduc-
tions in the turnover tax that had wrought the miracle of 1925. Total
indebtedness of public bodies rose, apart from the partial revaluation of
prewar and war debts, to 7.7 billions in 1927—1928 and by another
6 billions net2 till the end of 1929.
Analysis of the expenditure thus financed undoubtedly reveals
admirable cultural and social achievement, eminently productive of
economic and supereconomic values, in comparison with which the costs
might even be called moderate. In particular, there is much to be set,
in terms of beauty as well as in terms of welfare, against the desperate
financial position into which the big cities maneuvered themselves.
But neither that cultural aspect nor the various deficits as such are
pertinent to our subject. The important thing is the unavoidable infer-
ence that we have here a case of an excess of consumption by public
bodies, inducing excesses all over the economic system, withdrawing
capital from industry, or preventing its being built up — directly by
1The figures refer to what is, in Germany, technically described as Finanzbedarf.
They include items such as interest on the public debt, which it is usual to class as "mere
transferences." Taxes plus contributions to social insurance amounted respectively to
18.3 billions or 24.5 per cent in 1925 and 5.4 billions or 11.5 per cent in 1913. They
increased to 19.6 billions or 27.8 per cent in 1929. See for all figures, for example, Statis-
tisches Reichsamt, Finanzen und Steuern in In-und Ausland 1930, p. 548. The figures for
national income are highly controversial (see Sec. E).
2 Increase from Apr. 1, 1928, to Dec. 31, 1929, was 6.56 billion marks, but redemption of
prewar and war debts amounted to about 560 millions. Those debts in foreign currency
or in units guaranteed against depreciation (Festwert-und Valutaschulden), which were
contracted during the war and inflation, are excluded.
718 BUSINESS CYCLES
taxation,1 indirectly by the ensuing rise in costs — and another illustration
of the Doctrine of Spending. While it lasted, this process of impoverish-
ment produced what many Germans at the time quite well characterized
as a "prosperity of consumption" (Konsum-Konjunldur) which, con-
trasting so strikingly with underlying difficulties, superimposed itself
on the ordinary run of cycles and produced figures of output that almost
incurred the Dawes plan penalties. No doubt many fellow economists
will call that the most normal state of things imaginable and deny with
due epithets any suggestion to the effect that it would, in any case, have
had to end either in a breakdown or in "inflation" followed by a still
more vehement breakdown.
4. But the particular manner in which the breakdown actually came,
as well as the effect of that policy before it came, was determined by the
flow of foreign (and expatriated German2) balances. Data are not
entirely reliable, but the main facts about this capital import stand out
clearly enough.3 "Restriction" having been decreed, i.e., "inflation"
having been definitively renounced by the Reichsbank on Apr. 7, 1924,
both demand for foreign credits and readiness to grant them manifested
themselves almost immediately. Although prevented, until the autumn
of 1925, by the embargo on foreign lending from offering German secur-
ities for public subscription, English banks and bankers displayed willing-
ness to resume short-term relations with former German customers as
soon as — and before — the Dawes plan was carried, and even a ten-year
private loan was negotiated in August 1924 (Prudential Insurance —
North German Lloyd). American capital took nine-tenths of the loan
to Krupp soon after the flotation of the reparation loan. These trans-
1 The political structure also proved to be unequal to its task in that it reduced unpop-
ular taxes while keeping up those that injured the economic engine. For 1929 the total
load of taxation resting on industrial net returns, including everything, was about 90 per
cent in the highest brackets, in many cases more than that. See No. 4 of the monographs
(Einzelschriften) published by the Statistische Reichsamt. It does not follow, of course,
that without those taxes net returns would have equaled the figures from which the per-
centages were calculated.
2 Both during and after inflation, the flight of "capital" from Germany must have been
considerable, and part of these balances probably returned under foreign flags. Figures
are not available, but this matters little for our argument, since expatriated capital pre-
sumably behaved much as did bona fide foreign funds.
3 An official survey of foreign indebtedness was first made in 1926. After the break-
down, however, and when the "standstill agreements" were being negotiated, a more
comprehensive investigation — as of July 28, 1931 — was made, and some further data were
added later. The international commissions appointed in pursuance of the recommen-
dations of the London conference (July 21 to July 23, 1931) issued two reports, known
respectively as the Lay ton and the Beneduce reports (August and December 1931), which
presented most of the available material. Among the literature on this subject the reports
of a committee of the Verein fUr Sozialpolitik (Die Auslands-Kredite 1928, ed. W. Lotz,
Schriften vol. 174, Chap. Ill) should be particularly mentioned.
1919-1929 719
actions ushered in a long series of similar ones : states, provinces, munici-
palities, semipublic corporations, especially light and power concerns,
public and semipublic credit institutions, religious bodies of all types,
as well as banks and industrial companies all rushed, in spite of the
"suasion" and more drastic measures of the Reichsbank, for the oppor-
tunity of borrowing long in terms of foreign currency at what, counting
everything, on the average came to roughly 9 per cent.1
They did more than that, however. Partly because of that official
resistance to the rising tide of foreign indebtedness, partly for obvious
financial reasons, borrowers, especially concerns and banks, at the same
time borrowed short in whatever happened to be the most accessible of
the foreign money markets. These short loans were cheap only if
granted in foreign currency, but German debtors also owed, by the end of
July 1931, about 4 billion marks to foreign, mostly nonbanking, creditors,2
not all of whom were bona fide foreigners. The inflow of these various
funds was a function, not only of business, but also of political situations
and fluctuated considerably. But it persisted almost to the end of 1930.
By then the long-term foreign debt amounted, according to official
estimate, to the equivalent of 9.2 billion marks, and the short-term
foreign debt, after some repayments in consequence of the scare of
1929, to about 14.9 billions.3 Other foreign investments were estimated
at 6 billions. This total of, roughly, 30 billions provided the exchange,
first of all, for the transfers on reparation account equivalent to 10.3, and
for interest payments which amounted to 2.5 billion marks. It also
financed the building up of foreign balances and investments — 9.7 billions
and possibly more — and the net increase in Germany's gold reserve
(deducting the reduction in foreign exchange held by the Reichsbank)
which was 2.1 billions. Finally, it also covered the deficits in her com-
modity balance of trade, which during those seven years added up,
after deduction of 3 billions for "services rendered," to 3.3 billions.
This accounts for 27.9 out of those 30 billions. Various hypotheses
suggest themselves with regard to the difference. It is, however, not
1 The high cost to borrowers reflects, of course, primarily a large premium for risk.
But it must also be remembered that the cost of most of these transactions was very high.
Most of the securities placed in America, for instance — representing, roughly, two-thirds
of the total amount of the long-term foreign debt — had to be laboriously sold all over the
country to people of moderate means since, owing to the progressiveness of the American
income tax and to the tax privileges attaching to the Liberty loans, the higher return from
German bonds was not so attractive to large investors as one might think.
2 English banks had, since 1927, also granted some credits in marks to German banks.
These were not short, however.
8 The official survey mentioned in a previous note puts them at 12 billions, inclusive
of the 4 billions that were mark credits. But that figure is of end of July 1931; 2.9 billions
had been withdrawn in the preceding seven months.
720 BUSINESS CYCLES
greater than we should expect it to be with data of this nature, which
also leave many other points in doubt.
Thus, the excess of imports over exports absorbed only 11 per cent of
the total inflow of monetary "capital." Two-thirds of it — roughly, 20
billions — was used up by reparation payments and Germany's foreign
investment, including balances with foreign banks. It is not super-
fluous, however, to emphasize that the effects of this are inadequately
described by saying that reparations and foreign investments were
"paid out of foreign loans." This applies strictly only to a small part
of that sum, at most 2 or 3 billions, which were directly borrowed for
and applied to those purposes: only to this extent the economic life of
Germany was for the time being, in fact, relieved from all further effects,
exactly as if someone else had undertaken to carry those burdens for her.
The modus operandi of all the rest was more complicated. The sums
required were actually raised by taxation or paid by the individuals or
firms who wished to invest abroad, and all the foreign credits did was to
provide the exchange with which to transfer them. Schematically we
can represent this process by assuming that taxpayers and investors
bought the foreign exchange which borrowers had acquired, handing their
marks over to them and the foreign exchange to foreign governments or
sellers of assets. As thus financed, reparations and foreign investments
neither increased nor reduced available funds in Germany. The foreign
credits prevented temporarily all those adjustments of incomes, prices,
and the balance of commodity trade which otherwise would have resulted
from both. But they prevented them by a route that was different from
and much rougher than that of direct borrowing for reparations and
investments. This particular method was rendered possible by the
fact that, barring import surpluses and interest payments, German busi-
ness had no use for foreign exchange but needed marks for expenditure at
home. This — the salient — point stands out still more clearly with
respect to that amount by which foreign credits surpassed reparation,
foreign investment, import surplus, and interest requirements, and which
— partly reflected in the increase of the Reichsbank's gold holdings — of
course, swelled deposits.
The effects are clear: not only were adjustments to the facts of the
situation prevented, but the pulse of German business became dependent
on the rate of flow of foreign funds ; with foreign banks indirectly financing
a considerable part of investment and current operations in Germany, the
policy of her central bank was checkmated; the Konsum-Konjunktur
mentioned before was powerfully propelled; and of course, a financial
situation was created that was in constant danger of collapse on com-
paratively small provocation,1 especially, as it was hardly avoidable under
1 Many firms carrying on purely domestic business but having acquired the habit of
financing themselves by short foreign credits came, however prosperous their income
1919-1929
the circumstances that both German and foreign banks should, directly
and indirectly, knowingly and unwittingly, finance long-term industrial
commitments by short foreign funds. Thus, part of the foreign credits
effected precisely what an issue of greenbacks might have done: in a
sense, it camouflaged "inflation" by producing its results under the sur-
face of an apparently very "sound" monetary system.
But still more interesting than the effects are the causes of this practice
of financing Germany's domestic business in this way. Concerns bor-
rowed to such an extent because taxation absorbed what otherwise would
have become fixed and working capital. They borrowed abroad because
the same taxation — together with the preceding inflation — had reduced,
and kept on reducing, the means of savers, because public and semipublic
expenditure absorbed a large amount of the lending ("creating") power
of German financial institutions and because domestic credit was, in
consequence, both scarce and dear.1 The responsibility of taxation and
of public expenditure for the short credit situation, in particular, is
obvious. Assuming that, at the end of 1928, foreign short credits
amounted to 13 or 14 billions — an estimate that is arrived at by adding
50 per cent to the figure given in the Layton report — and deducting
(without increasing the figure, though this would almost certainly be
justified) German balances abroad, 4.5 billions, and credits for the current
financing of German foreign transactions, which were "revolving" and
not dangerous — this item, being very doubtful and to some extent also
overlapping with the first, we will put at not more than 3 billions — we
may perhaps estimate at a figure of the order of magnitude of 6 billions
the credits that eventually created trouble. How much of these does
the reader think would have flowed in at all if total public (not only
federal) expenditure had been kept at the 1925 level, if the surplus of that
year had been lent in the open market, if income and corporation taxes
had been reduced, and if interest rates and prices had behaved as in that
case they assuredly would have? It should be added that, though the
"consumers' prosperity" would then, no doubt, have failed to come
statement, in danger of bankruptcy whenever a cloud darkened the political sky. Since
short credits will always take to flight in such cases, there is no need of attributing such
phenomena to foreign, e.g., French, ill will. But the logic of the situation itself created a
sensitiveness of large sectors of business to erratic shocks, which was altogether abnormal
and cannot be explained on any of those principles that we ordinarily use in analyzing
economic fluctuations and "spirals."
1 It follows that taxation, in fact, made reparation transfers possible. Only, it did so
in a way very different from that which some economists thought of: they thought that
taxation would reduce system expenditure in Germany, thereby depress the price level and
thus produce the requisite export surplus. This it did not do precisely because of the
foreign credits. But it forced people to borrow abroad for domestic purposes, and so
produced the requisite foreign exchange by a different method, but for the time being not
less effectively.
722 BUSINESS CYCLES
about, it does not follow that consumers' welfare would have been, even
in the short run, substantially impaired.
If into this picture we insert the borrowing, both domestic and foreign,
of the municipalities and the fact that much of the expenditure even of
business concerns was unproductive in a commercial sense,1 we have
before us the groundwork of the theory of the specifically German pros-
perity of the later twenties, which is thus seen to link up with the socio-
political pattern of the time in more than one instructive way, and also
of the specifically German form of the subsequent economic breakdown,
as well as of much besides. Some traits will be added later; many more
would have to be added if complete analysis were feasible here. No
suggestion of a one-factor causation is intended. But our facts and their
consequences certainly suggest a lesson.
5. England's postwar situation may usefully be compared to her
situation in 1815. As has been stated in Chap. VI, the national debt was,
in relative importance, similar in both cases and so was the depreciation
of the currency and the burden of taxation. The cases differ however in
that the Napoleonic wars were times of vigorous industrial and commer-
cial expansion while the World War left England with her industrial
organism impaired and a loss of foreign investments which it took her
1 The question of "productiveness" — or the cognate question: What is to be classed as
an "investment"? — is a delicate one for other reasons besides the irritation it invariably
causes. It bears upon our subject in two respects. First, when Mr. Schacht during his
first tenure of office at the Reichsbank fought the rising tide of foreign debt, one of the
arguments was that only such loans should be permitted that would issue in investments
"productive of foreign exchange." Whatever might be urged against the correctness of
this argument, its real meaning, viz., that it was dangerous to finance by foreign borrowing
anything except commercially profitable enterprise, was under the circumstances undoubt-
edly much to the point. The municipalities, at which the Reichsbank's attack was chiefly
aimed, had no difficulty in replying that no less than 94 per cent of all their foreign loans
(as of Mar. 31, 1928, see O. Mulert in the volume of the Verein flir Sozialpolitik previously
quoted, p. 38) were in fact applied to directly paying propositions, electricity being the
biggest item. For municipalities very naturally approached the foreign investor with the
most businesslike-looking part of their spending programs and not with plans for resplendent
town halls, the funds for which were, nevertheless, set free by borrowing abroad for the
former. The productivity of, at all events, a very large part of the total indebtedness,
domestic plus foreign, both of public and semipublic bodies and of private concerns, is in
fact open to doubt. Second, there has been a tendency to define "investment" so exten-
sively as to make the term practically useless for our purpose. It is owing to this only that
the impressive figures of real investment were arrived at, which have been compiled by the
Institut ftir Konjunkturforschung (Sonderheft 22, 1930), and which at first sight seem to
contradict our diagnosis of impoverishment, excess consumption, and interference of tax-
ation with the "formation of capital." To begin with, investment within our meaning
of the term means economical investment (minimum outlay per unit of net return) : a
railroad is, no doubt, an economic enterprise, but not everything that is spent on it is
therefore an investment, however wonderful (and some German station buildings are not
less than wonderful) it may be. This applies even to expenditure conducive to excellence
1919-1929 723
over 10 years to make good.1 Conquests of political and economic
positions were made in both cases, but net gains were insignificant in the
recent case as compared to the earlier one. Also, practically the whole
world lay before England's industry and trade in 1815 and instances of
industries that had irreparably lost their previous positions, so con-
spicuous in postwar times (coal, cotton textiles, shipbuilding), were
then few, if not altogether absent, the only danger zone being agri-
culture— which then stood for the "key industries" that now had to be
"safeguarded."
In one respect, the frame of mind in which England encountered
postwar problems in 1918 was curiously like that in which she tackled
the situation of 1815. In both cases, return to the gold standard at par
was considered, by dominant opinion, as a matter of course. The report
of the Cunliffe committee,2 which was appointed as early as January
of service. That is also one of the reasons why part even of industrial investment must
be looked askance at from this standpoint. A large part of the "investment" in agri-
culture, especially that part which merely covered deficits, was clearly unproductive.
Moreover, public buildings do not constitute investments, any more than Louis XIV can
in any useful sense be said to have invested when he built the palace of Versailles. But
neither does housing unless the dwellings are let at rents which fully cover cost, the long-
term rate of interest included. More than anywhere else it is here necessary to ward off
misunderstanding by stating that the writer, if he thought his personal value-judgments
worth presenting, would have to confess that he thoroughly approved of every single one
of the, roughly, 11 billion marks that were spent on it from 1924 to 1928 (of which more than
half was from public sources, 4 billions being loans at 3 per cent financed by a housetax
levied from houses which inflation had freed from debt, the llauszinssteucr introduced by
the emergency decree of Feb. 14, 1924). This, however, is entirely irrelevant to our dis-
cussion. Relevant is merely that this was largely consumers' expenditure producing the
effects of consumers' expenditure. Thus corrected, the impressive total of net investment
amounting to about 39.5 billions, which the Berlin Institute presents for 1924 to 1928,
dwindles to considerably less than half. It should be added that increase in inventories
should be included only insofar as it is not due to inability to sell and that the figures for
internal accumulation of concerns (undivided profits) must be interpreted in the light of
the fact that there was a systematic downward bias in depreciation accounts, because,
after inflation, the new gold values were very generally put at figures which were too low.
The statement to be found in a recent study that two-thirds "of the [foreign] capital
supplies made available to the German economy in 1924 to 1928 • • • were applied to
extension*of plant and equipment" is hence seen to be misleading. A good discussion of
facts and problems, not quite however in accord with the views expressed in this note, as
well as most of the relevant literature and material, will be found in E. Welter, Die Ursachen
des Kapitalmangels in Deutschland, 1931.
1 Great Britain's foreign investments reached the prewar figure, roughly 4 billion
pounds, in 1931, but even then the presence of about 240 millions of short balances must
be set against it. Also a significant change had occurred in the character, particularly the
productive possibilities, of those investments.
2 The laconic Bradbury report (Committee on the Currency and Bank of England
Note Issue, 1925) brushed aside all questions of principle and can only be described as a
practical clincher.
724 BUSINESS CYCLES
1918, does indeed contain matter that would have been uncongenial to
some of the authors of the bullion report, but there is hardly a difference
in fundamental principle between the two documents. In recommending
speedy reestablishment of the gold standard at prewar parity — though
without free coinage for private account — to be achieved by the accumula-
tion of a gold reserve of 150 million pounds and the gradual reduction of
the amount of paper money in circulation, it expressed a view and an
intention which at that time undoubtedly prevailed. This % must be
accepted as a datum of the situation.1 In order to understand it and the
consequences it entailed, two things must be kept in mind. First, that
decision was fundamentally extrarational, and all arguments that have
been adduced for it, either from virtues of the gold standard in general —
largely imaginary under the circumstances — or from the particular
interests of England as the world's banker — which position was partly
untenable in any case and, moreover, as much endangered as buttressed by
that policy — were no more than ex post rationalizations of what really
was a foregone conclusion and to many minds involved the national
honor or, at least, prestige. Second, the public and especially those
labor men who favored that policy were certainly not aware of the
sacrifices — great or small — which in the short run it was bound to entail.
And those responsible men who may reasonably be assumed to have been
aware of them seem to have overlooked that 1918 was not 1816, i.e.,
the fact that their policy would have to work in an uncongenial social
environment. What the public wanted and what the " responsibles "
were driven back upon thus amounted to an attempt to swim the channel
without getting wet. Perhaps it is surprising that the impossibility of
playing the orthodox game just in one sector of national policy while
it was clearly up in every other should not have occurred to bankers and
1 By considering that volonte generate as a datum, the writer merely wishes to stress a
fact which is relevant to his narrative. But, as the text will amply show, he has no wish to
"justify" it. Since many eminent English economists feel so strongly on this and cognate
points, it will be conducive to better understanding to state expressly that the writer
largely agrees with the advice and criticism offered by some of them at the time and later,
in particular with the practical upshot of Mr. Hawtrey's arguments. As will also be seen,
he rates at a lower value both the effects of the monetary policy pursued and the possi-
bilities of available alternatives. And he does not think it irrelevant to speak of the
effects that policy might have had if other elements of the economic and social pattern
had shaped differently. But within a clausula rebus sic stantibus he believes Mr. Hawtrey
to have been, in the main, right as far as practical policy is concerned. The opportunity
may be taken to add an analogous statement with respect to other matters of England's
economic policy, which have to be touched in passing. In particular, there is not only no
intention to attack the policies recommended by a brilliant group of English fellow econo-
mists, but iri1 many if not in all cases the writer entertains no doubt whatever about the
wisdom, ex visit of English short-run interests, of the advice proffered by them. This is
perfectly compatible with great differences in theory and diagnosis.
1919-1929 725
politicians. Perhaps it is not. But in any case the ultimate failure
must be understood in this light and not in the light of any general merits
or demerits of such a policy, which are an altogether different matter.
Actual "deflation" (in the sense of reduction of the amount of cir-
culating medium without a corresponding reduction in physical volume
of transactions) was attempted but, meeting with resolute resistance to
the necessary adjustments, was given up quickly. All the government
did beyond liquidating war expenditure was to reduce the circulation of
currency notes.1 The Bank continued its efforts to accumulate gold.
Its stock of coin and bullion fell slightly in 1922 but almost reached the
Cunliffe goal by 1925 (at the end of the year, 144.6 millions) and surpassed
it in 1926 (at the end of the year, 151.12 millions). Notes in circulation
roughly moved in step until 1927. Other Securities were close to their
1919 figure both in 1924 and 1925. Other Deposits fluctuated strongly
around a level that declined but little from 1919 to the end of 1921.
Total Clearings fell in 1921, when Deposits of London Clearing Banks were
still increasing, and began to recover toward the end of 1923 (see below,
Sec. F). This, for the moment, is all the background we want. It also
suffices to assist the reader in forming an opinion how much the monetary
element as such can possibly have had to do with the crisis of 1921 and
the slowness of the recovery in the two subsequent years. So far the
analogy with the course of events from 1815 to 1821 stands out suggestively.
Meanwhile, the pound was practically left to follow its course. On
the cessation of American pegging, it had begun to fall, discount reaching
its high point, 34 J^ per cent, in February 1920.2 But it speedily recov-
ered and hovered around a 10 per cent discount in the summer of 1924.
This recovery cannot to any great extent have been due to the actual
measures (excluding speeches) taken by government or the Bank. Nor
can it be fully accounted for by the fall in price level that had occurred
but in the main was an international phenomenon. Much more impor-
tant was that same factor, which subsequently also made the last step
so temptingly easy: by that time the whole world was expecting that
England would return to the prewar parity and, hence, was .buying
sterling exchange. The temptation proved irresistible: the hope that
gold would internationally fall to meet the pound was, to say the least,
vague; the domestic situation offered but little ground for optimism; real
success in the sense of achieving a gold standard that would normally work
1 There were 367.6 million pounds of them in 1920 and 295.6 millions in 1925. Later
measures dealing with them need not be discussed in this book.
2 An official embargo was laid on gold and silver, but it merely continued a state of
things which, by patriotic discipline, had obtained throughout the war. After that official
embargo had been lifted, attempts were made amidst the difficulties that ensued to resort
again to the latter method by frowning on gold exports, mobilizing public opinion against
operators, inducing shipping companies to raise freights for gold, and so on.
726 BUSINESS CYCLES
at prewar parity was for the moment out of the question; but technical
success was within easy reach. Short balances flocked to England
to profit from a rise that was a practical certainty, and by December 1924
the pound was no more than 1J^ per cent below parity. The only diffi-
culty was in the dangerous pressure which was to be expected the very
moment parity would be reached. For then speculation would naturally
realize and withdraw. Against this danger several defenses were built up.
An American credit constituted one safeguard to be used if necessary
after the event. The second safeguard was a bank rate of 5 per cent,
which the Bank was careful to make effective.1 The third consisted in
keeping the pound down until as nearly as possible the moment of the
plunge. This was successfully done by buying up foreign exchange,
ostensibly for the purpose of servicing the debt to the United States.
Once we accept the goal, nothing but admiration can be felt for what
once more was a very fine piece of steering, that not only achieved
technical success but, in doing so, also avoided jerks and jolts and mini-
mized injury to the economic organism. Reasons will be offered (Sec. F)
for believing that the Bank's policy continued to deserve the latter com-
pliment during the subsequent years. By adroit use of gold devices and
masterly handling of the short-loan market it undoubtedly made the best
of very delicate situations. It is nonetheless clear that the Gold Standard
Act marks not the end but the beginning of the real difficulties. On the
face of it, the new gold standard was untenable and bound to break
down — if worked according to classical principles — through an unavoid-
able efflux of gold. Obviously, the Bank and the government must have
hoped that it would be possible for a time to maneuver so as to avoid the
latter and that during that time domestic or international developments
would resolve the dilemma. The first proved, in fact, possible, the latter
•failed to mature. It is at this point that the difference begins between
the course of events after the Napoleonic war and the course of events
after the World War of our time. Then, drastic readjustment, by means
of monetary policy, of the price level and of incomes would have been
possible, but vigorous evolution of the national economy made it unneces-
sary. Now, no comparable development setting in, drastic readjustment
by monetary policy would have been necessary, but it was not possible.
The very fact that substantially similar monetary policies produced such
different results and that hardly any difficulties of a purely monetary
kind were encountered in the one case, while difficulties of this nature
eventually proved insuperable in the other, should convince anyone that
monetary policy has no claim in a general theory to the key position
1 Great funding operations coupled with the reissue of about the same amount of
treasury bills achieved that end. On the other hand, banks were made to understand that
foreign lending was not for the time being considered as in the national interest ("embargo
on capital export").
1919-1929 727
allocated to it by some economists. It is no contradiction — but, on the
contrary, a trivial corollary — to say that if all the other elements of the
social and economic setup are taken as given and if we are left with
monetary policy as the only variable, the latter will acquire a causative
importance and, in particular, become, if incompatible with the rest of
the setup, a depressing external factor. This is what happened in our
case and what defines the role of England's monetary policy in her post-
war cyclical process. Overvaluation of the pound — putting a bounty on
imports and penalizing exports — and a bank rate abnormally high under
the circumstances are two familiar instances of its modus operandi,
although we should not overestimate the importance of either.
Thus our argument leads, in this sense, to agreement with those
English authorities on money who hold that, all other things being as
they were, the return to the gold standard or the return to it at prewar
parity spelled pressure that aggravated difficulties and could have been
alleviated by another monetary policy which in turn need not have
produced other difficulties. Much more important for explanation,
however, although under the circumstances perhaps not for practical
advice, were some of those other things. Neither monetary policy nor
the fact that England's domestic and international position was not so
favorable in 1918 as it was in 1815 will sufficiently account for what is
universally felt to be unsatisfactory economic performance. The funda-
mental social change, which must destroy the frame and atmosphere
conducive to the working of the capitalist engine at maximum efficiency
and which we have tried to analyze in Sec. B, is, with the exception of
Russia, nowhere so obvious as it is in England. The difference between
what Parliament did with the income tax respectively in 1816 and in 1918
is a significant symptom even for those who refuse to look upon it as a
cause. England's fiscal policy1 characterizes a social situation which
hardly displays any symptoms that might be interpreted in the opposite
sense.
What is typically English in this, however, and what strikes any
outside observer more than anything else is the unrevolutionary form
of a change which, among other things, involved as great a transfer of
wealth as was ever effected by any revolution, the Russian one alone
excepted. By this the writer does not intend to refer only to the absence
of a violent break of legal continuity but to the much more relevant fact
1 As we have seen in Chap. VII, the roots of that policy reach far back. Let us recall,
however, that the earliest conspicuous landmark in a long development was the first
budget of the Campbell-Bannerman administration (Asquith; Mr. Lloyd George's "people's
budget" was the second major step). The essential point was the earmarking of part of
the surplus that resulted from Boer War taxation for the purpose of financing old-age
pensions instead of applying it to the reduction of taxation or debt, which would have been
the "classical" measure to take.
728 BUSINESS CYCLES
of continuity in the personnel that mans the political ship. The socio-
logical pattern of that personnel changed, no doubt, significantly. But
it did so slowly. And what at any time was the old stratum succeeded in
absorbing new elements both in the sense that it readily received and in
the sense that it effectively assimilated rising talent. This was possible
only because the old stratum itself — or its more active elements — had
an altogether unique ability to accept and to handle fundamentally new
situations and principles. The set of people for whom ruling is more
important than the purposes and interests to be served by ruling is in
England larger and more influential than anywhere else. That set
carried free trade in corn, although it was at the time controlled by
agrarian interests, and through generations managed both domestic and
foreign policy entirely from the standpoint of a bourgeoisie, with which
it was more or less allied but by no means identical. There is, thus,
nothing — short of very great inconvenience — to stop it from repeating
the feat and running a Labor party or even, without doing this, running
the country by means of a Conservative party on Labor lines, scaling off,
no doubt, rough edges but not altering the fundamental contour of events.
To a certain extent this is what actually happened in the postwar period,1
and the fact that England emerged from the war in a substantially
conservative mood — which was intensified, if anything, by the social
unrest, the outbreaks in England, and the revolution in Ireland in 1920 —
and subsequently kept Conservative administrations in power for most of
the time, can hence not affect our diagnosis or its implications for the
mechanism or the results of the cyclical process. Again, nothing the
Labor party actually did, either in office or out of it, can with any con-
fidence be pointed to as a major cause of disturbance deflecting the eco-
nomic process from the course it would otherwise have taken. Its short
spells of power and its strong position in the country are relevant for us
merely as indications of the changes that have occurred in social structure
and atmosphere. In financial matters, in particular, it was eminently
"sound'* in the orthodox sense.
1 Many Englishmen will not agree to, and even be irritated by, the above statements.
But this disagreement is easily accounted for by the difference in standpoint between the
man in the thick of the game and the mere observer. For the one, the individual measure
matters in all its details, and in 6ghting for his points he will use, and have to use, a stereo-
typed phraseology that knows no colors except black and white, not less but more so when
the actual colors shade off into each other. For the other, nothing matters except broad
results and very simplified contours. Also, Englishmen who struggle with the less mobile
elements of their own environment will, no doubt, be unable to see what strikes the writer
as extreme adaptability to new situations and principles, if principles be the right word for
attitudes or opinions that are so easily jettisoned without sterile regrets. Yet that adapta-
bility is a provable as well as most important fact. Peel and Disraeli were masters of an
art that has, obviously, not been lost,
1919-1929 729
War financing, too, had been on classical lines. The writer finds it
difficult to understand the critics who were not satisfied with what to him
seems admirable performance. The postwar budget of the coalition
government has, however, repeatedly been criticized on the score of —
epithet to be inserted by the reader, e.g.t meritorious, nefarious, bene-
ficial, abominable — extravagance. This is relevant to our subject
because it shows that rapidity of liquidation of war expenditure can
hardly be invoked in explanation of the subsequent slump, except in the
sense that continuance of war expenditure at an increasing rate would
have avoided it for a time. Subsequently, expenditure was normalized,
but on a level about four times as high as that of 1913-1914. The excess
profits tax was abolished. Of other adjustments we need mention only
the reduction in the flat rate of the income tax from 72 pence in the
pound, in 1918-1919, to 48 in 1925-1926. Glancing at the state of things
in the latter year we find that expenditure of the government as per
Finance Accounts (which include transfers to local authorities and pay-
ments to North Ireland, but not other local expenditure) was 826.1 mil-
lion pounds and revenue 812.1, deficit turning into surplus if we deduct
the 50 millions applied to the reduction of debt. This compares to
about 198 million pounds, at which accounts roughly balanced for 1913-
1914. Revenue of local authorities from rates was 166.1 million pounds
in the later fiscal year, as compared with 79 in the earlier. Opinions
differ widely as to the choice of an index by which to reduce those figures
to comparability. It is sufficient, however, to state that while total
money income of the inhabitants of Great Britain and North Ireland
about doubled between the two years, total public revenue increased to
about 3.4 times its former amount, which, looked at from the standpoint
of intact capitalism, was already extremely high. Unlike German
public spending, that of England, especially in its permanent elements,
displayed some stability for several years afterward and never outran
revenue to any serious extent. The Consolidated Revenue and Expendi-
ture Account of government, local authorities, and social insurance
always showed a surplus.1
Fiscal policy undoubtedly interfered with the saving-investment
process, however.2 This effect becomes still more obvious if we take
1 Cf. C. Clark, National Income and Outlay, Table 59, pp. 140 and 141.
2 According to Professor Bowley and Sir Josiah Stamp, The National Income 1924
(1927, p. 57), out of an Aggregate Income of 2,020 million pounds in 1911, 320 millions were
saved and 225 paid in rates and taxes; while out of the 1924 aggregate of 4,165 millions
475 were saved and 855 paid in rates and taxes. Very roughly, the amount "spent freely"
increased as Aggregate Income, i.e., doubled; the amount saved increased by one-half, i.e.,
less than price level; the amount paid in rates and taxes increased to more than 8% times
its former figure. We may differ, of course, as to the effects on the economic process of
this absolute fall in real saving and this relative fall in monetary saving. But it would be
730 BUSINESS CYCLES
account of the fact that the burden of taxation was not only increased
but also shifted in a way that cannot have failed to affect saving and
accumulation. Neglecting the excess-profits duty and the tax on cor-
porate business, taxation of inheritance and income yielded about 74.5
millions in 1913-1914 and about 389 in 1925-1926, and though con-
sumers' expenditure of the payers of these taxes was certainly curtailed —
nobody at all familiar with English life can have any doubt about that —
the greater part, perhaps two-thirds of the difference of £.40 million
pounds, which we arrive at if we double the prewar figure in order to
make it roughly comparable, must have come out of potential savings or,
in some cases, out of dissavings. The implications of this — see above,
sub 2 and passim — may be distasteful. All that can be replied is that
it is really a pity that facts have a way of verifying views which are so
obviously antiquated. For the symptoms we simultaneously observe in
the economic organism are exactly what old-fashioned economists (whose
theories the writer, as this book amply proves, is in general very far from
sharing) would have expected to follow from such a fiscal policy.
Those implications are not weakened, first, by the fact that relatively
moderate but not negligible amounts were spent on subsidies to business
— agriculture, air transport, the merchant marine, coal mining, the beet
sugar and other industries. Nor, in the second place, is it relevant that
other lines of expenditure will command the support of most of us and
that still others certainly "increased productivity." Policies of social
betterment, for example, accounted for about 16 million pounds in 1912—
1913 and for about 72 in 1925-1926. * Also, it must not be forgotten
that improvements of the environment, however beneficial, do in most
cases fail to link up with the tax that finances them in such a way as to
neutralize its economic effects. There is, for instance, no doubt that if
government action involving expenditure could reduce precipitation in
the Lake District, the nation's enjoyment of that delightful country
would be greatly increased. But a general tax levied for the purpose
would act as a net burden all the same, and the benefit would simply
enter into the general environmental conditions exactly as if, instead of
H. M. Servants, Jupiter Pluvius had wrought the change gratis.
highly unreasonable to deny that this fall itself was mainly due to the increase in taxation.
The use of other estimates could not affect the result substantially. We shall, however,
see reason to suspect (see infra, Sec. F, II, 1) that all estimates greatly exaggerate the
amount of saving, which actually was far smaller than 475 millions.
1 Those are estimates of the Statistische Reichsamt, Staatsausgaben von Grossbritan-
nien, Frankreich, Belgien und Italien 1927. A subjective element enters into the various
definitions between which one must choose. Besides, the different economic situations
entailing — also for reasons that have nothing to do with taxation — more unemployment in
the one case than in the other must be taken into account. For our argument all this
matters but little. We are not criticizing.
1919-1929 731
Finally, third, it must, in view of a not very creditable discussion that
has arisen about the point, be emphasized that taxation for the service
of the 5.9 billions by which the domestic national debt had been increased
in consequence of the war, cannot be left out of account on the ground
that it effects a "mere transfer." It is true that an internal debt in
some respects raises problems which differ from those presented by an
external one. It is also true that this type of expenditure does not reduce
the amount of factors of production available for industry, as expenditure
on additional policemen would in a case of full employment. But
exactly in the same sense — though in no other — in which expenditure on
disabled soldiers is a burden, expenditure on disabled capital is, too.
Its effects on those who pay the corresponding taxes are the same as the
effects of taxes levied for any other purpose. And no gains of either
class of recipients balance this loss: both the disabled soldiers and the
owners of the "capital'* which financed the war live, without con-
tributing, on the results of the productive efforts of the rest of the com-
munity, while otherwise they would have been earning wages and interest
by increasing the national dividend. No talk about "putting money
from one pocket into the other" avails against this fact.
Other effects apart, this fiscal policy enforced recourse to foreign
short balances in a way that presents some analogy with the German case.
The sober management of the budget prevented, indeed, a "consumers'
prosperity" in the German sense from developing; and part of the
English short borrowing links up with long-term lending abroad. But
it still remains true that if taxation had not cut so deeply into potential
savings, less foreign balances would have flowed in and many things,
money rates among them, would have been different from what they were.
However, the English situation, as pointed out above, more than that
of any other country was influenced by circumstances affecting particular
spots. These circumstances, such as the innovations which depressed
coal mining, or the rise of native capitalism in India, which depressed
the cotton textile industry, or the losses England's international position
suffered in the fields of banking, shipping, and insurance, permanently
altered the conditions of her economic life and were for her largely external
factors. But they are much more important in the interpretation of
her cyclical phenomena than whether banks buy more or less assets.
Moreover, both her domestic and her international environment were
bound to change at an increasing rate in the neomercantilist age. This
in itself unavoidably imposed a radical change of policy. Cobden himself
could not be accused of inconsistency — although, of course, he might still
be wrong — if he now rose from the dead to preach the doctrine that pro-
tection, state enterprise, and managed money must be resorted to in order
to transform the Empire into an ironclad world of its own.
732 BUSINESS CYCLES
D. Outlines of Economic History from 1919 to 1929. — A very rough
sketch will be sufficient to convince the reader that all the major features
of economic life during that period in fact conform closely to our idea of
a Kondratieff downgrade and that none of them fights against the hypo-
thesis which this turn of phrase implies.
1. We begin with the agrarian sphere. Both preceding Kondratieff s
displayed within their negative phases prolonged agrarian depressions.
We have seen that in causation and symptoms they differed sufficiently,
as between each other and in each case as between countries, to cast
doubt on any very broad generalization about them, particularly with
respect to the "necessity" or "normality" of their occurrence. We have
also seen, however, that certain properties of Kondratieff downgrades
tend to produce depressive conditions in the agrarian world as a whole,
and that agricultural innovations, if any, tend to produce in sectors of
that world depressive conditions that may be important enough to
create a picture of general agrarian depression. Obviously, this is what
we find in the postwar period and what provides the first approximation
into which it is easy to fit all the other factors of agrarian situations.
But the latter are, nevertheless, important and should not be neglected
merely for the sake of one-factor theories and one-remedy therapeutics.
Primarily, the fall in agrarian prices was a fall not in relative but in
absolute price, i.e., an element of the fall in the general price level. Such
a fall is part of the mechanism of cyclical downgrades — of Kondratieff
downgrades, in particular. As we have seen, it would not, in itself,
suffice to produce an agricultural crisis, although it may adversely affect
the welfare of the agrarian community if the farm prices of products
fall more than the retail prices of the finished products which it buys.1
"Crisis'* may ensue, however, if the fall of the price level impinges on a
debt situation that has developed from borrowing either for unproductive
purposes — such as the acquisition of land — or for insufficiently productive
ones — such as mere expansion. But in the United States and England
agriculture had to face, as it had after the Napoleonic wars and, in this
1 This, of course, is invariably the case. As to farm machinery, there is an investigation
by the Federal Trade Commission (1920), which found that prices of farm implements
were advanced by producers and dealers in 1917 and 1918 by more than was "warranted by
the increase in their costs," especially in those lines which were practically controlled by one
concern — mowers and binders, for instance — and on which the "premium for innovation"
was more completely collected than on others. But it is true in general, that, partly owing
to local dealers* margins and to freight, which together amount in some cases to about 25
per cent of the price paid by the farmer, partly to monopolistically competitive situations,
prices of these things tend to be both high and rigid. Servicing, which the farmer cannot
as a rule do himself, is also expensive. The same applies to most of the gadgets of modern
life which the farmer buys as a consumer and, to a lesser extent, to finished industrial products
generally, although some of these are supplied to him efficiently and cheaply and without
any abnormal profit.
1919-1929 733
country, the Civil War, not that kind of fall in price level which is a
normal element of the 'economic process in Kondratieff recessions and
depressions, but the much more violent reaction of prices to the rise dur-
ing the World War. Moreover, agriculture had been, to the extent
indicated in the last section of Chap. VII, an innovating industry, or
rather an industry that had innovations forced upon it which originated
elsewhere, such as the internal-combustion engine, specifically agri-
cultural machinery, electric power and appliances, new fertilizers. As
we should expect, these innovations fully conquered and came to fruition
in the downgrade, and they, as well as the locational shifts, which con-
stitute the most important of agriculture's own innovations, sectionally
reduced costs to a level on which large sectors were unable to compete:
the food problem of humanity was, as far as the economic process was
concerned, indeed definitively solved, but at the expense of the agricul-
tural interest. Competition by other countries, development of which
was accelerated by the war, harvests, conditions of demand,1 international
barriers, and other factors have to be inserted, however, to complete the
picture as it unfolded itself from year to year.
a. Elaborating a little for each of our countries, we will first notice
that in the United States the Bureau of Agricultural Economics index
of prices received for farm products rose from 1915 to 1919 by 109 per
cent, while the index of commodities bought by farmers rose, until 1920,
by 94 per cent.2 The year 1920 brought a moderate fall and 1921 a
fall to 116 per cent of the prewar figure, from which the index of farm
prices recovered quickly, to reach a peak of 147 per cent in 1925. Then it
1 The above refers to the facts that demand tended to become inelastic in some countries,
particularly in the United States, and that the impulse that had caused it to shift favorably
all through the prewar time slackened in others (not in the United States). Also, new
modes of life engendering new tastes and habits tended to reduce consumption of the heavier
foods per head. But these facts have nothing to do with any shortage of either real or
monetary purchasing power of consumers. This, as the course of wages in our three
countries amply proves, is again — still more obviously than it was in the two other instances
— a pure myth, except in the cases of Germany and Austria during their inflations. Pro-
tectionist policies have more claim to our attention.
2 These indices are reproduced, for example, by E. G. Nourse in Recent Economic
Changes, vol. II, 1929, p. 548. Newer and more extended investigations have not greatly
altered the general picture. It is hardly necessary to point out the limits of the reliability
of such indices, particularly of the nation-wide type. Since interest on mortgages, rail-
road fares, and other items on which farmers' receipts are spent did not move at all, or
moved less than the index (taxes on farm property were in 1919, according to the same
source, only 30 per cent above 1914), the case in 1919 was still more favorable than it
seems to be at first sight. This only accentuates the subsequent reversal. Taxes on farm
property rose within our period to over two and a half times the prewar figures, and the
index of prices of the commodities bought by farmers never fell below about 150 per cent
of those figures. Discontent was, therefore, understandable, especially if we take account
of the rapid increase in wealth in the industrial sector. But discontent is not crisis.
734 BUSINESS CYCLES
fluctuated on a moderately falling "trend" up to the eve of the world
crisis. This development must be correlated with the development of the
agricultural debt. l Even in the (predominantly prosperous) prewar years
total farm mortgages were considerable — 3.3 billions in 1910.2 They rose
to 237 per cent of that sum by 1920, quite enough to produce many unten-
able situations, even if we take into account the fact that incomes had
risen more than farm prices, and afterward fell less than they did. But to
1925 there was a further increase to about 9.36 billions, the peak that
occurred in 1928 being but insignificantly higher — 9.46. Now, some
part of this load probably was the result of funding short-time debts
which had become irksome, and a greater part, the result of expansion
and of the mechanization to be mentioned presently. But the correlation
of the two periods of increase with rising land values is obvious, and the
inference is unavoidable that much of this increase in debt came from
purchases of land with a view to reaping not harvests but increments of
value.3 So far, then, we conclude: there was a short and sharp crisis in
agriculture in 1920—1921, which was part of the general postwar slump,
though accentuated for agriculture through the burden of partly unpro-
ductive debt. In the years to 1926 there was, however unsatisfactory
the situation may have been from other standpoints than ours, no general
agrarian depression at all. After 1926 and to the threshold of the world
crisis, the agrarian situation became increasingly unsatisfactory, but the
only general cause of this was, again, the pressure of unproductive debt.
But this diagnosis misses, besides many minor points, a major one,
viz,, the influence of the innovations mentioned before. Some of them,
like the progress in the cultivation of citrus fruit and vegetables or in
refrigeration and canning, did not — or not materially — spell competition
of some sectors or products with others, and brought a net addition to the
1 Mortgages were, of course, only part of the total agrarian debt. But since part of the
short debts were — in cattle ranching, for instance — really for the purposes of current busi-
ness and not dangerous, we can confine ourselves, in a sketch of this kind, to the mortgages.
2 Figures on indebtedness are taken from the publication of the Bureau of Foreign and
Domestic Commerce on Long-term Debts, 1937, p. 107.
8 Official publications are understandably careful in dealing with that delicate point.
Still, the publication quoted in the preceding note goes so far as to say (p. 106) : "The war
and postwar period of rising land values and enlarged volume of land transfers resulted in
a marked increase in the volume of farm mortgage debt, as mortgages were freely used to
facilitate sales." In a country in which everyone speculated, there should be no shame-
facedness about this. Remedies are not found by concealing the patient's ills. Moreover,
although in some states that "free use of mortgages" went rather far, it is not held that the
speculation in farm land in the country as a whole went to anything like the lengths it did,
for instance, in New Zealand. In Mississippi, 83.7 per cent of farms were mortgaged; in
West Virginia, only 32.6 per cent. The point under discussion has been officially recog-
nized. See, for example, Secretary H. A. Wallace's statements in The Agricultural Situa-
tion, issued by the Bureau of Agricultural Economics, June 1, 1937.
1919-1929 735
total of agricultural incomes. To a lesser extent this is true also of
poultry and cattle rearing and of dairying. 1 Others, like some electrical
appliances, even helped sectors that were being put out of existence by
competition, especially those whose main difficulty was dear labor. But
most of the improvements in the methods of agriculture, while instru-
mental in bringing forth agriculture's share in that rising tide of con-
sumers' goods which, according to our schema, is a feature of Kondratieff
downgrades, and even producing agricultural prosperity in wide sectors
of the country,2 tended to push certain regions below the margin of
profitable production. This, of course, is wholly true of the productive
success achieved on reclaimed, drained, irrigated lands and of the process
by which large areas have been taken into intensive cultivation of crops
and into horticulture that previously served purposes of "extensive"
farming. But it is partly true also of the truck, the tractor, most of the
machinery newly introduced into grain farming, and to some extent of
the use of electrical power. Most of them increase the optimum size
of the farming unit, some of them can be used to full advantage only
under the particular conditions of the Great Plains. From 1920 to 1930
the number of motor trucks increased from 139,000 to 900,000 and the
number of tractors from 246,000 to 920,000.3 The latter invites the
combination of operations that were previously quite distinct, ploughing
and the preparation of the seedbed, for instance, and thus steadily leads to
ever-increasing mechanization. The use of the combine harvester, which
had first been a success in California, spread and yearly sales increased
nearly sevenfold in the same period. Cotton- and corn-harvesting imple-
ments must be added, but no further examples are necessary in order to
establish our point. Nothing of all this was fundamentally new; all of
it is typically "induced development" of the kind which on previous
1 An eminent politician, instead, professed himself shocked at finding that Wisconsin
milk and milk products are being sold in Georgia; but neither farmers nor economists are
likely to share the feelings he obviously entertains on the subject of national division of
labor.
2 That prosperity was, however, less accentuated than might have been expected, not
only because profits had to be shared, as pointed out before, with the industries that were
responsible for the innovations, but also because of the perfectly competitive character of
agriculture, which responded to the lowering of costs by a prompt reduction of the prices
of products. This is what, together with the fact that the undersold units did not promptly
disappear, created that impression which is sometimes conveyed by the phrase "agri-
cultural overproduction." If the implication is that farmers are to "blame" for their
plight, the reply is in order that, by behaving as they did, they fully rendered the social
service that circumstances enabled them to render. In any case, however, that phrase is
misleading, for the state of things envisaged was not simply overproduction in the ordinary
sense of the word.
3 The low-priced tractor (Fordson, International Harvester), in particular, did not come
in before 1915.
736 BUSINESS CYCLES
occasions we found to be characteristic of Kondratieff downgrades. Our
old formula, Depression spotted by Prosperity, fits the case as it did the
others.1 Emigration from agriculture to industry was, under the circum-
stances, from the standpoint of the logic of the capitalist mechanism, a
perfectly normal phenomenon of adaptation.
Other aspects of the same features and some additional features of
the postwar agrarian situation will come into view if we glance for a
moment at cotton and wheat in particular.2 Ever since the beginning
of the nineties the price of cotton moved fairly well with any all-com-
modity index, domestic consumption of cotton — also roughly — with any
index of industrial production. This on the whole remained so in our
period, the chief exception being the rapid recovery of cotton consumption
in 1921, right from the beginning of the year. Quantity exported was
below the average of the last prewar years in 1922 to 1924, but roughly
on the same level, or somewhat above, in 1924 to 1929, value rising
sharply from 1921 to 1925 and receding afterward. Rayon was only one
of the competing commodities that must have exerted some influence —
with increasing wealth, the competition of wool increases in many lines —
but owing to the emergence of new uses, such competition was of but
minor importance; the standard fiber was still to come. There was the
migration to lands made available by new methods of cultivation,
especially to Texas and Oklahoma, partly due to the tractor and the
mechanical picker (complemented by a corresponding innovation in
ginning), with a consequent competitive annihilation of much of the
Southeastern cotton farming.
In all this our process shows to perfection, and the process of labor
being drawn away from an "old" stratum (toward the Northeastern
industry) is particularly in evidence. Farmers' price of the standard
xlt follows, of course, that "price parities" between agricultural and nonagricultural
products are irrelevant for an appraisal of the degree of welfare enjoyed by the agricultural
community. But "income parities," as between the agricultural community as a whole
and the industrial sector, have not much more economic meaning. There is not more
reason to expect agricultural prices or incomes to keep a certain relation to other prices or
incomes than there is to expect that, say, coal and income derived from coal mining should
permanently keep its relative position in an economic world that is incessantly being
revolutionized. If for extra-economic reasons agriculture is to be subsidized — and there
are many such reasons — it should be done without reference to any such parities between
the present and a bygone state of things. But total market value of farm products, while
it fell much more than money income of industrial workers in 1920, from that year on sub-
stantially kept pace with it. See M. Ezekiel, Evaluating 1933 for the Farmer, Journal
of the American Statistical Association for June 1934, chart on p. 140.
2 The same phenomena would stand out still better in other instances, such as rubber,
coffee, and rice, which are, however, not products or not leading products of any of our
countries. The studies on the Artificial Control of Raw Material Supply published in the
London and Cambridge Economic Service (J. W. F. Rowe) should be mentioned.
1919-1929 73T
quality rose from about 12^ cents to about 28^ cents during the war
(the latter figure being of November 1918), which was perfectly normal
and neither justified nor actually induced increase in acreage. Acreage
harvested actually fell from its 1914-1915 peak. The ravages of the
boll weevil in 1921, 1922, and 1923, however, raised it to 32 cents toward
the end of the latter year, and this presumably propelled the expansion
in the West, which in spite of abandonments — not all due to the boll ,
weevil — had set in before and carried total acreage in this country from
the 29.7 million acres in 1921 to about 45.8 in 1926. Acreage outside the
United States at the same time increased from about 28.5 to over 42 mil-
lion acres, not only, of course, in response to that price, but in consequence
of the endeavor made in many countries to develop cotton growing,
which date far back into prewar times and were indirectly fostered by
the tariff policy of the United States. Thus there developed slowly,
beneath the surface of current fluctuations, an untenable situation which
was bound to curtail the role of American cotton in the world and to
explode in a major depression. The presence of overproduction in the
proper sense of the term is, in this special case, as undeniable as the
rationale of the argument for planned retreat.1 Cottonseed oil and its
residues cannot be dealt with here, but their possibilities in the fields of
human and animal food and of chemical products, although very con-
siderable, would not fundamentally alter the picture.
The postwar wheat situation presents fundamentally the same fea-
tures, yet differs in important respects. Before the war, United States
production had indeed met increasing competition in the world's market
— from Canada and Argentina, in particular — but effects were always
compensated by favorable shifts in demand. After the war, this was no
longer so. Although population increased strongly, consumption per
head did not. On the contrary the latter decreased considerably in
1 We are speaking, be it remembered, of the pre-crisis cotton problem, and the above is
intended to submit that even before and independently of the world crisis there was a case
for what we call planned retreat. It should be clearly understood that this means two
different things. First, locational and technological innovation in the ordinary course of
the cyclical process of evolution spelled distress for a great part of the Eastern cotton-
growing industry. In a competitive industry dominated by types which do not in ration-
ality and promptitude of action answer to capitalistic requirements, it is, of course, possible
to argue for planning or relief in order to facilitate transitions, which in themselves are normal
and unavoidable but which the people themselves are unable to effect, and to mitigate the
hardships of the "competing-down process." But there was, in the second place, another
case for planning, viz., for planned reaction to planned expansion abroad. It is the latter
that constitutes the special feature in the postwar cotton situation. Whether this country
was to enter into competition with the new sources of supply opened up abroad or to retire
from the international market is not under discussion here; either course would have
involved planning by public authority in order to be successful. Into the implications of
this for a theory of planning we cannot enter.
738 BUSINESS CYCLES
response to changing tastes and habits,1 though the increase in the former
was sufficient to increase total consumption, which at the end of our
period was about 15 per cent above the average of the last five prewar
years. Foreign demand rapidly fell from its war peak after the cessation
of American war and emergency credits, both foreign competition and
protection accounting for the sharply falling "trend" and, from 1926 on,
the uninterrupted fall in quantity exported. World production, exclud-
ing Russia and China, after having decreased from 1915 to 1917^, increased
to 1928 by more than one-third and then moved about a level approxi-
mately 20 per cent above the last prewar years.2 European production
alone, including Russian exports, more than recovered in strongly inverse
covariation with United States exports. Interpretation of these facts
must, moreover, take into account the very low elasticity of domestic
demand. Some economists hold that production adapted itself to the
new conditions and point to the sharp decrease in acreage harvested per
head of population that occurred from 1919 to 1925.3 Production had,
however, expanded considerably from 1915 to 1919 (the trough of 1917
was due to the failure of winter wheat) and though contraction followed
fairly promptly upon the fall in prices, persisting excess capacity and that
inelasticity of domestic demand would nevertheless account for strong
effects of the fall in exports and of the variations in harvests.4 It does
hot follow that, because the price of wheat moved very much as any
all-commodity index, conditions peculiar to the wheat-growing industry
had nothing to do with it. The argument for planned retreat thus again
suggests itself.
But the central fact was the technological revolution.6 The average
yearly product during our decade of roughly 850 million bushels of wheat
may not look formidable in itself. But it was not the result of harmon-
1 But Domestic Disappearance of Flour, as calculated by the Stanford Food Research
Institute (see Wheat Studies for December 1932, p. 310), fully kept pace with the postwar
percentage of population growth. For grains in general the gradual elimination of horses
and mules must also be taken into account.
2 For Argentina, Australia, and Canada, taken together, increase as compared with the
last prewar years was over 40 per cent. Excepting Argentina, there was not much increase
in average yield per acre.
3 See, for example, Warren and Pearson, Prices, Chart 42 on page 50.
4 Only for 1918, 1919, and 1924 did good harvests spell good business for farmers. The
poor crops of 1920 and 1921 helped to steady prices, and the decade's minimum crop (1925)
brought what may be termed recovery values. But the increasingly good harvests of 1926,
1927, and 1928 exerted a pressure unknown to prewar times and led up to the collapse of
1929.
5 This has, of course, never been entirely overlooked. But the paramount importance
of that element of the case has been first pointed out by E. Altschul and F. Strauss, Tech-
nological Progress and Agricultural Depression, National Bureau of Economic Research,
Bulletin 67, 1937.
1919-1929 739
ious expansion in all parts of the country, which it would have been
possible to restrict again at proportionate and moderate sacrifice for every
grower of wheat or which, in fact, would have restricted itself without
catastrophe in the course of a few years of depression. It was the net
result of spectacular expansion in some regions and painful elimination
in others. Expansion was general up to 1919, even the East and South
responding to war prices. But the really significant increase in acreage
was not. That was confined to Montana, Kansas, Nebraska, Texas,
and a few other states and, obviously, was not due simply to war con-
ditions. Similarly, decrease in wheat acreage from 1919 to 1925 was
general but also unequal, hardly any decrease occurring in Montana,
for example. The subsequent expansion to 1929 coincided with restric-
tion in the South and East, where acreage decreased by about one-fifth
for the decade. Diagnosis of this course of events is obvious. Expan-
sion was in the Great Plains, where the mechanized farm, the tractor and
the combine thresher in particular, can be worked to full advantage and
yield acceptable returns at a price of 60 cents per bushel or less.1 Con-
traction was enforced where those innovations were not profitable and a
price of one dollar per bushel covers cost only on the better soils. We
recognize all the features embodied in our model and especially the
"competing-down process," passing sentence of economic death on
perhaps half of all wheat farmers. The implications of this do not con-
cern us here.2 But it must be added that this component of the postwar
situation, in fact, originated, as it should, in the preceding Kondratieff
prosperity. The great per-cent increase in Montana, Kansas, Nebraska,
and Texas was from 1900 to 1915. We may even go so far as to say
that this is what can be attributed to innovation per se, while the further
rise to 1919 — roughly, 10 million acres in the Great Plains — was a war
effect. Innovation would have spread and taken full effect in the down-
grade, as it always does. A depressive situation would, hence, have
ensued in any case. But war prices and reaction to them accentuated it,
which is all that prices or monetary factors have had to do with it.
b. For England and Wales, total value of marketed products of
domestic agriculture was, in 1925, 225 million pounds, less than one-third
of consumption. Two-thirds of that amount was value of animal prod-
ucts; less than one-fourth, value of cereals and potatoes. Total acreage
in cultivation had not increased at all during the war — it had slightly
fallen — and moved below the average of 1871-1875 throughout our
decade. Prices, of course, behaved much as they did in the United
1 See Mr. C. L. Holmes' report on Farm Production Costs as affected by Mechanical
Farm Equipment, United States Department of Agriculture, 1931.
2 That state of things, of course, accounts for the low average of farmers' income per
head.
740 BUSINESS CYCLES
States; relative prices of animal products recovered strongly from 1924
on.1 Income from agriculture, as per income tax statistics — which are
not very reliable in the case of agriculture — rose to 1920-1921 (54 million
pounds) fell to 1922-1923 (20.6 millions) and was then fairly stable (at
about 24 millions) until 1929. In the absence of a dangerous load of
debt, no untenable condition resulted. Investment that would not have
paid was at a low level. But the medium-sized farm bore up well on the
whole. It is obvious that no significant effect on the national situation
can have been exerted. Neither was there any "crisis" until 1929-1930.
There was, of course, the boom in the subsidized sugar-beet production,
quantity harvested increasing from 56,000 tons in 1922-1923 to 1,472,300
tons in 1927-1928.
c. In Germany, inflation had reduced the agricultural debt from
17.5 billion gold marks (1913) to about 2.7, the amount eventually result-
ing from the revaluation of mortgages.2 But it quickly mounted up
again. By June 30, 1926 the debt newly incurred — mortgage and other,
but not counting current debts with retailers and so on — amounted to
3.7 and by Mar. 31, 1930, to T.66 billions. If to the latter figure we add
2 billions for unreported floating debts and the revaluation mortgages,
we arrive for the end of our period at something like 12 or 13 billions,
and at an interest charge of over 1 billion. In order to understand this,
we must bear in mind the extent to which German agriculture and its
equipment had suffered during the war. Considerable investment was
necessary to restore it to normal efficiency. The inducement to run into
debt was stronger for the larger estates, particularly in the East, because
they suffered more from higher wages and taxes,3 because they had more
opportunity for trying out new methods involving investment, because
their owners' standard of life proved less adaptable than that of the
peasantry, and because it was primarily to them that expensive loans
were offered from public and private sources with a readiness that proved
1 There was, in spite of some increase in pasture, no significant increase in livestock.
But dairy products and poultry did increase; 82 per cent of the butter consumed and more
than half of the meat consumed in 1924 was imported, however.
2 No account need be taken of the Rentenmarkgrundschuld, i.e., the amount of 2 billion
marks for which German land had been mortgaged by law in order to provide a guaranty
for the Rentenmarks issued in the process of monetary stabilization. Agriculture had to
pay 5 per cent interest on this, but there was really no capital debt. When the Rentenbank
was liquidated (act of Aug. 30, 1924) it was succeeded (act of July 18, 1925) by the Benten-
bankkreditanstalt, which took over its assets and served as an intermediary agency for the
loans to agriculture that came from public and foreign sources.
3 Taxes on agricultural land rose more in the West than in the East, viz., from 8.8 marks
per nectar in 1912-1914 to 37.6 marks per hectar in 1924-1926 in the West, and from 7 to
25.3 marks in the East. But net returns shaped much more unfavorably in the latter
region than they did in the former.
1919-1929 741
fatal. Over 78 per cent of all holdings were eventually in debt in the
West and almost 90 per cent in the East.
If we are to trust the figures presented in the official Survey of German
Economic Conditions (Wirtschaftsenquete) for 294 medium-sized and
large units, net returns per hectar, which were 93 marks in the average of
1912-1914, were practically wiped out in the average of 1924-1926.
Another semiofficial source1 gives 3 marks for 1924—1925 and 18 for
1925-1926. These, of course, were the very worst years until 1930 to
1932, and net returns recovered somewhat in the interval. Nor do such
figures disprove the impression that a large part of the peasants, especially
those in thickly populated sectors of the country, did quite well. But
it seems impossible to avoid the inference that in many cases returns
barely covered, or even failed to cover, taxes and interest throughout the
period. As a result, foreclosures (as measured by acreage, not by num-
ber) rose above prewar level by 1926 and then steadily increased to a
maximum of 193 1.2 This does not even tell the whole tale, because
creditors may often have been prevented from foreclosing by the knowl-
edge that bidders would not be forthcoming and because there were very
many desperate situations which yet stopped short of technical insol-
vency. Prices of agricultural land were, it is true, even in 1929-1930
above prewar level,3 but only for holdings below 20 hectar. For larger
ones they were below that level, the more so, the larger the holding.
Diagnosis, though entirely different from that of the American and
English cases, is not difficult. The technological component plays only
a minor part. According to the census of 1925, German agriculture was
then but imperfectly mechanized, even in the class of units above 100 hec-
tar. The use of sowing, planting, and harvesting machinery was, of
course, fairly widespread among the latter, but only a minority used
steam and motor ploughs, for example, and the majority of peasants did
practically without any modern machinery at all. Only 644,713 of the
5.1 million units used electromotors and the tractor had not yet invaded to
a significant extent the domain of the horse,4 the ox, or the cow. There
was, indeed, great progress in fertilizing. Average yield per hectar was,
nevertheless, for all grains lower in 1924-1928 than it had been in 1911-
1 See Veroffentlichungen des Deutschen Landwirtschaftsrates, Heft 16. The data refer
to the income statements of those units that reported to the Accounting Office (Buchstelle)
of that Council of Agriculture, according to a set schema.
2 After that, foreclosures were drastically reduced by various government measures.
3 See W. Rothkegel, Die Entwicklung der Kauf-und Pachtpreise fUr LandgUter und
StUckl'andereien in the publication of the FriedrichList Gesellschaft on Deutsche Agrarpolitik
I.
4 There were in Germany (postwar territory) 8.8 million horses in 1913 and 3.7 in 1928.
The use of horses in agriculture, hence, must have increased, if anything.
742 BUSINESS CYCLES
1913, and land in agricultural use had decreased.1 Germany's dogged
effort toward self-sufficiency was to come later. It must not be for-
gotten, however, that our comparison is with the results of the innova-
tions of the Kondratieff prosperity, during which yield per hectar had
greatly increased, and that we deal with national averages only, which
veil significant differences between different regions. But such further
innovation or spread of previous innovation as there was in our period
hardly affected noninnovating sectors as it did in this country, because
the dominant factor of the German situation was foreign competition:
all domestic innovation did, under the circumstances, was to ease things
for some regions or individual holdings.
For the same reason, the changes in consumers' habits mattered less
than one might suppose. But while demand for meat, dairy products,
fruit, and vegetables developed satisfactorily from the beginning of the
period, there was a similar — though less pronounced — reduction in the
consumption of breadstuff s per head, as there was in the United States,
coupled with a shift from rye to wheat, which accounts for the (small)
increase in wheat and the (greater) decrease in rye acreage,2 the latter
indicating the depressed conditions which, with the exception of 1927 and
1928, prevailed in the most important branch of Germany's production
of grains. Those effects of a rising standard of life must also have
influenced potato growing. Conditions varied greatly in different parts
of the country, and the statistics of direct consumption and of the use
of potatoes in hog and cattle rearing and as an industrial raw material
present difficulties into which we cannot enter. The general picture is,
however, perfectly clear. A technologically very rigid supply, which had
developed in prewar times, proved too much for the market of a com-
modity that was in the process of losing its position as a staple food.
The ensuing "surplus" — aided by a similar surplus of rye — partly
explains the rapidity with which the hog population recovered from the
massacre inflicted upon it as a war measure: in 1927 it surpassed its pre-
war figure. This and competition from the East and Southeast brought
down the price of pork, so that all the most important products of Ger-
1 From 1918 (postwar territory) to 1927 by about 300,000 hectars. But the figure
should not be implicitly trusted. That decrease may be due to difference in methods of
surveying. There was certainly no significant increase, however. Throughout this and
the following chapter, the term "postwar territory" refers, unless otherwise stated, to the
German territory as delimited by the Peace of Versailles.
2 The rye acreage fell from 5.2 million hectars in the average of 1911-1918 (postwar
territory) to 4.7 millions in the average of 1926-1928, while the wheat acreage rose from
1.66 to 1.69 millions. Conditions in wheat growing were relatively favorable throughout
the period. Value of the harvest surpassed the average of 1908-1909 to 1913-1914, which
was 721.5 million marks, by 1925-1926, when it was 791.2 millions. Cf. H. Paetzmann,
Zur Lage der Deutschen Landwirtschaft, Vierteljahrshefte zur Konjunkturforschung, 1926,
Sonderheft No. 3, p. 18.
1919-1929 743
man agriculture eventually failed, except locally, to cover costs and
burdens.
Thus, we have before us the case of an industry which, if left to itself
and other things being equal, would have had to contract ever since the
late nineties. Protection and successful innovation enabled it to hold
its own and even to expand during the Kondratieff prosperity. But in the
downgrade the superiority of American opportunities — giving much
larger scope, in particular, to mechanized farming — and the cheap labor
of the countries to the East and Southeast of Germany were bound to
assert themselves. It then became clear that her agriculture was unequal
to servicing its rapidly growing debt and to the wage rates and burdens
of the modern state — that its marginal value-productivity was hopelessly
below the general national level. It is from this angle that the plight of
German agriculture must be understood. The state of things could not
be inferred from the behavior of prices, which were favorable in some
cases, especially in those of milk and butter, and not obviously catas-
trophic in any case. The upsetting factors were on the side of costs.
Recognizing this yet wishing to avoid the laisser-faire conclusion from
this, public policy exhausted the arsenal of protection, prohibition (at
first in the form of trade monopolies), reduction of railroad rates, direct
and indirect subsidies, and also resorted to "internal colonization,"
breaking up a number of larger estates into small settlements. These
measures and further borrowing kept things going until both gave way
before the onset of the world crisis.
2. Next, the postwar building booms call for comment. Of their
quantitative importance in the economic processes of the period it is very
difficult to give an exact idea, but very easy to give one that is approxi-
mate. If, for example, we accept the statement contained in the 1929
Census of Construction, that building will on the average (directly)
give a year's full employment to one man for roughly each $5,800 spent,
and if we take into account the employment created by the production
and transportation of building material (which we might now attempt
to do from the data of the Federal Employment Stabilization Board and
the Bureau of the Census) and in other subsidiary industries and, by
way of secondary effects, in all industries, we cannot doubt that con-
struction was the chief contributor1 to the postwar business volume in
this country as well as in the other two. This is no more unexpected than
1 As will presently become clear, this does not mean that there was a one-way relation
between building and the rest of the economic organism, or that the prosperities of the
period originated in the building trade. No mere appeal to the quantitative importance
of construction within the total of system expenditure would have any explanatory value.
In most cases it is obvious that construction rather responded to than created conditions
favorable to its expansion. The latter was the case only insofar as there was innovation
in the building industry itself.
744 BUSINESS CYCLES
the postwar agrarian depression. Building booms, in particular booms
in residential, public, and public utility construction, occurred in the
downgrades of both preceding Kondratieffs — for instance, in England
in the twenties of the nineteenth century, in all three countries before
1873, in the United States from 1878 to 1894. All of them, with one
exception,1 were stronger than any that occurred during Kondratieff
prosperities.
Nor is this a mere matter of history. Taking, for brevity's sake,
dwelling-house building only, we need but list the factors that would
produce supernormal activity, in order to see that the general conditions
prevailing in Kondratieff downgrades and revivals — more precisely, in the
prosperity phases of the shorter cycles which run their course within
Kondratieff downgrades and revivals — are more favorable to the occur-
rence of building booms than are the general conditions prevailing in
prosperities. Falling rate of interest is one of them. High rate of
increase in real incomes is another: from rising or constant money incomes
of the middle and lower classes, accompanied by falling cost of living,
new demand for better housing will naturally follow. Innovation in the
building industry or its subsidiaries will work in the same direction
because, like other innovations, it is likely to spread in recession. The
rise in rents that occurs during Kondratieff prosperities supplies, barring
a subsequent fall in money incomes — which, as we have seen, is not
likely to occur — an additional stimulus. Finally, industrial evolution in
general means industrial migration and, moreover, migration from the
countryside to the cities, both of which create new demand for construc-
tion that is eventually provided for during recession. Of course there
were, besides, other factors, unconnected with the features of the Kondra-
tieff phase which happened to prevail. The omissions of the war period,
both as to replacement and as to normal increment, constitute the most
important of them.
a. In the United States the war did not interfere with either residential
or other construction to anything like the same extent as in England or
Germany, but such indications as we have leave no doubt about the fact
that it was at an abnormally low level in 1917 and 1918.2 At least
in most parts of the country, a short-lived boom set in at the end of 1918,
during which building costs rose sharply — by 25 per cent or more. This
was followed by a drastic fall both in building activity and costs, and
from the beginning of the fourth quarter of 1923 the postwar boom in
residential building definitely got under sail. There was a setback in the
1 That exception is the English building boom from 1895 to 1905. It set in during
revival, but it persisted through nearly half of the subsequent Kondratieff prosperity.
2 Cf. J. R. Riggleman, Building Cycles in the United States, 1875-1932, Journal of the
American Statistical Association^ June 1933.
1919-1929 745
second and third quarters of 1924; then a peak was reached in 1925,
descent from which lasted to about May 1927; another peak occurred in
April 1928. After that we have decline, which though at varying rates —
in 1930 there was some retardation — continued to February 1933, with
1929 and 1931 displaying the sharpest falls. In apartment -house and
hotel construction the maximum occurs in 1926, but the figure for 1925
comes near it and those for 1927 and 1928 are not much below it.1
Expenditure on new, nonfarm, residential building, including hotels and
clubs, is estimated by the National Bureau of Economic Research at
34 billion dollars for the decade.
Diagnosis of that boom, which was entirely financed from private
sources, presents no difficulties. At the beginning of the period there was
dammed-up demand. Population, in spite of the Immigration Restric-
tion Act of May 1921, increased from 1920 to 1929 by 15 millions, the
largest absolute amount of increase per year in the history of the country.
1 On the sources of construction statistics and some of the pitfalls they contain see C.
Gill, Construction Statistics, Journal of the American Statistical Association, March 1933.
The above refers to values of contracts awarded and is mainly based on the data of the F.
W. Dodge Corporation, which until 1932 exclude contracts of a value less than $5,000 and
— an advantage from our standpoint — alterations and repairs. The data, steadily increas-
ing in accuracy, specialization, and coverage, start in 1901, cover 27 states from 1922, 36
from 1923, and later on 37, from which the national total is arrived at by a multiplier. The
writer wishes to acknowledge the F. W. Dodge Corporation's courtesy in answering his
questions and permitting use of material. No other series of contracts awarded have been
used except to test the writer's guesses as to particular questions. Building permits are
available (Carl Snyder's series) for seven cities since 1882. Of other permit series, the
annual one published by the Bureau of Labor Statistics — on account of its specialized
information — and the monthly one published by the Federal Reserve Board have been most
useful. But a report by the National Bureau of Economic Research, foreshadowing the
results of a comprehensive investigation working with permits (D. L. Wickens and R. R.
Foster, Non-farm Residential Construction, 1920-1936, Bulletin 65, Sept. 15, 1937) seems,
on the one hand, entitled to so much confidence and, on the other hand, to yield results so
materially different that the writer is now inclined to doubt the1 reliability of his own work
in the matter summarized in the text. In particular, the second peak (1928) may be
untrustworthy. Instead, new residential construction is held by the authors of that report
to have declined steadily from the first peak (1925). But there is no difference as to the
abruptness of the fall in 1929 or the steepness of the ascent in 1925. Physical volume (there
are now, however, new estimates of numbers of units constructed, e.g., L. J. Chawner's in
the Annals of the American Academy of Political and Social Science, vol. 190) must be derived
from those value figures by deflating. Several indices have been constructed for this
purpose. They include, however, labor and materials only. The one used by the writer,
published by the Federal Reserve Bank of New York, attributes (constant) weights of
respectively 45 and 55 per cent to those two items (weighting according to the 1929 Census
of Construction: labor 33 materials 47 per cent, other items 20 per cent). Professor Mills
(Economic Tendencies in the United States, p. 267) rightly deflates the dollar volume of
each type of building by an index appropriate to it. The reader is referred to his treatment
of the subject. But from 1925 to 1929 all cost indices change so little that the general
picture is affected by deflation only for 1919 to 1924.
746 BUSINESS CYCLES
There was also considerable internal migration. Real income per head,
rising strongly in all brackets, made that demand effective and added
new sources. The motor was the only other "expensive" one of the
items toward which the surplus turned. From 1916 to 1920 rents had
risen on a national average by almost two- thirds. They fell but insig-
nificantly, even in the crisis of 1921. Primarily the boom was a response
to these conditions. Building costs rose swiftly in 1923 and after 1924
remained fairly stable on a somewhat lower level.1 Interest on urban
mortgages was, though falling,2 not particularly cheap as compared with
other long-term rates, except where building was financed by bond issues.
But under the circumstances of that period and in the glow of its uncritical
optimism neither costs nor interest charges mattered much. It seemed
more important to get quickly the home one wanted — or the skyscraper
the prospective rents of which in any case compared favorably with the
rate on mortgage bonds — than to bother whether it would cost a few
thousand dollars — or in the case of the skyscraper, a million or so — more
or less, provided money was readily forthcoming at those rates. And
it was. First mortgages on urban real estate represent, on the one hand,
not all the loans that were made available for building and, on the other
hand, also financed not only other types of building but other things than
building. But it is still permissible to point to the fact that they increased
from, roughly, 13 billions in 1922 to, roughly, 27 in 1929, building and
loan associations contributing about 7.8, commercial banks about 5.2,
mutual saving banks 5.1, life insurance companies 4.8, and mortgage
bonds more than 4.3 This increase is out of all proportion, not only with
the increase in what can in any reasonable sense be called savings, but
also with the expansion of bank credit in other lines of business, and
illustrates well how a cheap money policy may affect other sectors than
those in which it is conspicuously successful in bringing down rates.
1 The increase in building costs in 1923 is beyond doubt. The decrease that occurred
in 1924 was a few per cent only and figures out differently with different indices. No
substantial reduction occurred before 1931, at least as far as quoted prices are concerned.
2 It is interesting to note that the beginning of the boom — like that of the English boom
of 1982 (see infra, Chap. XV, Sec. D) — preceded such fall as there was in mortgage rates.
In this as in other cases it is perfectly clear that other factors are more important than
interest, although our description sufficiently shows that the writer has no intention to deny
its influence. Residential building fell off, however, at the very time bond yield began
to fall more markedly.
8 Those figures are taken from C. E. Persons, Credit Expansion, 1920-1929, and its
Lessons, Quarterly Journal of Economics, November 1930. They differ somewhat from
those of J. H. Gray and G. W. Terborgh, First Mortgages in Urban Real Estate, Report
for the Real Estate Research Committee of the Brookings Institution, 1929. Both esti-
mates exclude second mortgages and other items. Professor Irving Fisher, Booms and
Depressions, 1932, p. 173, puts the total nonfarm mortgage debt in 1929 as high as 37 billion
dollars(?).
1919-1929 747
If such a sector display a very elastic demand for the funds which that
policy will drive toward it, interest in it need fall but little or not at all in
order to produce all the consequences that we usually associate with
"too-low" money rates.
Innovation lent its aid. The steel-skeleton structure, made cheaper
by steadily increasing use of reinforced concrete and workable by the
electric elevator, had created new possibilities ever since the nineties, and
these possibilities had become realities — and a feature of — the Kondra-
tieff upswing. In the downgrade after the war this innovation, improved
by several minor and "induced" ones, propelled by changes in the habits
of life that made the apartment increasingly desirable to the American
bourgeois family and by the plethora of credit, spread and conquered,
much like motorcars or rayon and exactly like those innovations that
carried the prosperities, and spread in the downgrades, of preceding
Kondratieffs. Similarly, prefabrication, primarily made possible by
the use of the new materials but also applied to stone and lumber,
extended its domain far beyond the skyscraper. Excavation of base-
ments by means of power shovels improved by the caterpillar tread and
belt and bucket conveyors, the use of power hoists, of power concrete
and mortar mixers, and of pneumatic riveting machines, rapidly became
a matter of course for contractors in all lines of building1 — typical down-
grade developments, all of them. Their full effect — the mass production
of the perfectly standardized and mechanized cheap house — is still to
come, however. During our period the ordinary family house was in
the main still being built in substantially old-fashioned ways by small
and inefficient firms.
But the conclusion that this essentially consequential development —
in response to the omissions of the war period and increased real pur-
chasing power of all classes, on the one hand, and in response to previous
innovation, on the other — issued in overbuilding, owing to the additional
stimulus imparted by the monetary factor, must not be accepted hastily,
however plausible it may seem. Some types of response to those con-
ditions, especially the ones that were linked to speculative real estate
operations, were clearly of the bubble class. The Miami case may serve
as an example. Nor can there be any doubt about the merits of the
financial methods that were also used in less "speculative" cases — New
York skyscrapers for instance — and in particular about the financial
quality of the mortgage bonds, which increased from 682 millions in 1922
to 4,169 millions in 1929, and which were readily lent against by banks.
Finally, everything was done to make it easy for everyone to run into
debt, for the purpose of building a home as for any other purpose. It is
indeed, easy to understand that such a structure would give way, not
1 See Jerome, Mechanization in Industry, pp. 134-145.
748 BUSINESS CYCLES
only under the impact of a serious crisis, but even in consequence of a
mere failure of rosy expectations about things in general to come true.
In other words, we shall readily understand why the load of debt thus
lightheartedly incurred by people who foresaw nothing but booms
should become a serious matter whenever incomes fell, and that con-
struction would then contribute, directly and through the effects on the
credit structure of impaired values of real estate, as much to a depression
as it had contributed to the preceding booms. Nothing is so likely to
produce cumulative depressive processes as such commitments of a vast
number of households to an overhead financed to a great extent by com-
mercial banks. But this does not quite amount to saying that there was
overbuilding in the sense that the amount of construction was greater
than it was possible to absorb without losses under the conditions then
prevailing, and that this excess was an independent cause of the great
depression.
Rents fell from 1924 on,1 but only moderately. Vacancies increased,
but not more than was to be expected in a period of rapid obsolescence of
existing house property. The big, old, ugly, and inconvenient house
soon became difficult to sell, because of changes in tastes — some of them
attributable to the automobile — and because of the increasing wages
and decreasing efficiency of servants. But there is no reason to believe —
in particular, the fact that the increase in dwelling units was greater
than the net increase in married couples does not prove — that the spurt of
1925 could not have settled down into an appropriate average activity
and that even the results of speculative excesses could not have been
liquidated without any violent crisis in building, let alone in general
business. As a matter of fact, this was accomplished to a certain extent.
If we accept the figures of the National Bureau of Economic Research,
we arrive at the conclusion that four years of such adjustment — including
local crises — actually followed upon that boom without much general
disturbance being created. In the final result, expansion in this line was
not so obviously greater than it was in other lines of consumption that
explanation of subsequent vicissitudes could simply be given in terms of
"malin vestment." And incomes had first to fall because of a general
crisis, for the special crisis of building and of real estate to come about.2
1 All indices agree in this; there is, however, some doubt whether or not the national
average temporarily increased again somewhat in 1926 and 1927.
* The reader is welcome to argue, if he feel so inclined, that the above comes to saying
that all that would have been necessary in order to avert serious trouble would have been
continued increase of incomes by means of, say, additional government expenditure. It
is more important, as well as more useful, to point out that timely restriction was not the
only alternative open under the circumstances. Much could have been achieved by devis-
ing more appropriate methods of financing by public or semipublic institutions, which
might have kept commercial banks out of this business and created new forms of loan
contracts less sensitive to the impact of depressions.
1919-1929 749
This analysis refers to residential building only. Results are not,
however, substantially changed by including other types. One of them,
commercial building, is perhaps still more than apartment houses and
hotels exposed to the suspicion of speculative overdoing. Contracts
awarded1 increased steadily to a peak in 1927 and another almost as high
in 1929, and summed up for 1922 to 1929 to nearly 6.7 billions, the rate of
increase over the period being substantially in excess of that of residential
building. Industrial building increased at a still greater rate — contracts
awarded sum up to about 4.8 billion — but there is very little reason to
suspect any excess over the requirements of the general march of things.
Unlike the other items, but also conforming to expectation, this moved
well in the Kitchins and showed equally well the sweep of the two incom-
plete Juglars.2 Finally, more than one-third of the grand total3 of con-
tracts awarded — nearly 49 billions according to the Dodge figures, which
certainly understate — comes under the heading of Public, Institutional,
and Utility construction. Part of this is reflected by the increase in
municipal bonded debts. Alone federal expenditure on new construc-
tion, repairs, and alterations amounted, from 1920 to 1929, according to
the Federal Employment Stabilization Board, to about 2.5 billions, the
trend being upward all the time and 1929 displaying the highest figure
(308 millions4) — a fact worth mentioning in view of the prevalent talk
about insufficient spending. According to the same source, the figures
of which are again incomparable with those used above, the expenditure
of railroads — steam and electric — and power and telephone companies
on construction and maintenance moved, from 1923 to 1929, extremely
steadily on a slightly rising level, summing up to 20.4 billions. But these
sums, of the importance of which those examples suffice to give an idea,
were expended in ways that would not produce any material effect on the
economic process beyond what is implied in the expenditure itself.
b. In Germany privately financed construction during the war was
probably at less than one-fifth of the 1910 to 1913 average and in the end
practically ceased. Some impulse was given to it by the "flight into
1 According to the Dodge figures of value of contracts awarded. They must not be
added to or compared with the National Bureau's estimates of residential building. For
this purpose they would have to be increased, though presumably not as much as the Dodge
estimates of dollar volume of residential building.
2 The figure for 1923 was almost 80 per cent above that of 1922; 1924 shows decrease;
1925, over 80 per cent above 1924, indicates the rise of the fourth Juglar, which starts in
the second half of that year; 1926 reflects prosperity in the act of gathering momentum;
1927 is influenced by the Kitchin depression; 1928 shows recovery; and 1929 brings the
peak of the period (845 millions).
8 It is worth while noticing that that total which reached its peak in 1928 was very
stable from year to year, 1925 to 1928, and fell by 12 per cent in 1929.
4 This expenditure then rose to new levels in 1930, 1931, and 1932. For this and the
following statement, see C. Gill, op. cit., pp. 39 and 40,
750 BUSINESS CYCLES
real values" during the inflation period, but this impulse ceased to act
in the time of "wild" inflation. From 1924 to 1929 industrial and com-
mercial building presents a picture not unlike the American one. It rose
from 1.23 billion marks in 1925 to 2.16 in 1925, fell in 1926, more than
recovered in 1927, and reached its peak of nearly 3 billions in 1928,
when commercial and industrial contracts awarded in the United States,
not yet at their peak however, amounted to about 6.7 billion marks.
Though the latter figure understates actual expenditure to an unknown
extent, it is nevertheless clear that, considering the difference in the
size of the respective economic organisms, Germany did in this respect
not much worse than this country. There was a small fall already in
1929, but for 1930, the further fall being but 11 per cent, the comparison
is still'more favorable. The total from 1924 to 1929 was 13.45 billion
marks, about 1 billion greater than the total for public building — includ-
ing roads and canals — which behaved similarly but fell off much more
than business building in 1930. 1 Residential building was subsidized,
as has been stated before. It is doubtful whether it would have revived
of itself under the circumstances, since the rigorous control of rents of old
dwellings absorbed part of the demand and made people unwilling to
pay rents which would have covered costs. As it was, great building
activity set in, expenditure steadily increasing to a peak in 1928 —
followed by a considerable drop — of 3.4 billion marks, which compares
with about 3 billion dollars in the American peak year, 1925, or with 4.7 if
we accept the National Bureau estimate: the comparison, however
precarious, because residential building does not cover exactly the same
things in both countries and for other reasons, is yet not without sig-
nificance considering the difference in real income per head. The grand
total for the period was 40 billions, to which residential building con-
tributed 14 billions. Unlike the corresponding American amount, this
sum largely failed, independently of any crisis, to give economic return.2
1 The madness of this kind of reaction to the world crisis — for madness is the mildest
term the writer can think of when placing himself on the standpoint of the gospel of spend-
ing— ceases, however, to be so obvious when the internal and external political situation
and the previous fiscal policy are properly taken into account, see the following chapter.
We should note here that, especially in the last years of the period, the financing, par-
ticularly of municipal building, was the last word in " unsoundness " and may compete with
anything skyscraper financiers did in this country. In some cases such things as bridges
were financed by six months' paper. The total may also have been understated, for some
items were tucked away in municipal budgets of extraordinary reticence.
2 The Survey of Economic Conditions (Ausschuss zur Untersuchung der Erzeugungs
und Absatzbedingungen der deutschen Wirtschart. Third subcommittee, Der Deutsche
Wohnungsbau, 1931, p. 18) records the opinion that economic calculation was put out of
operation (ausser Kraft gesetzt) in that field. Although intended to be cautiously reserved
and perhaps the result of a compromise, the passage is misleading and even unjust. As
pointed out, there was reckless financing and there may have been some waste. But
1919-1929 751
c. It was in August 1932 that the great English building boom began
that was to solve the housing problem in the same sense in which it can be
said that the development of the textile industries in the three decades
following the Napoleonic wars solved the clothing problem and the
development of agriculture after the World War solved the food problem
of the masses — leaving many things to be done, no doubt, but only things
of the second order of difficulty and importance.1 That was the housing
boom which corresponds to the American one in the twenties. The
building activity during the period under discussion presents essentially
different features. Our remarks will, however, be confined to residential
building also in this case.
The general conditions of a Kondratieff downgrade would, ceteris
paribus, have been as favorable in England as they were elsewhere to the
occurrence of a housing boom. In 1912 and 1913 such a boom actually
seemed about to develop in continuation of the one that started in the
nineties. And after the war there was, of course, dammed-up demand;
the number of marriages rose considerably, if temporarily, above the
1913 figure and so on. But other things were not equal, and one feature
of Kondratieff downgrades failed for the time being to materialize:
at least to 1924, per capita real income was not much higher2 than it
had been 10 years earlier. Hampered also by high costs, building there-
fore showed little tendency to revive in 1919. While in this respect the
English was dissimilar to the American case, it was dissimilar to the
German case, not only in the resources that were available to cope with
the situation, but also in the sober spirit in which it was met. This
spirit asserted itself in three ways. First, rent control, imposed during
and continued after the war, was administered in a much more business-
like way than in Germany. Second, private enterprise — especially
large-scale private enterprise — was harnessed to the purpose rather than
discouraged, and a workable mixture of public and private activity was
experimentally found. One of the contributions of this policy to the
boom of the thirties was precisely that it trained the building industry to
its task, while at the same time the appetite of people who could afford
to pay for better housing was being stimulated. Third, it was recognized
in fact — whatever actual motives and phraseologies may have been — that
the road toward satisfactory housing of the lowest strata led through
extensive dwelling-house building was, under the general conditions of the time, the Kond-
ratieff phase included, a perfectly normal phenomenon, and a public contribution toward
the creation of satisfactory homes need not have created financial trouble.
1 This is obvious in the case of food and clothing, and for housing conclusively proved
by the Ministry of Health report on overcrowding in England and Wales, 1986.
2 According to Professor Bowley, it was, if anything, somewhat smaller. The question
will be discussed in Sec. E.
752 BUSINESS CYCLES
providing houses for that large number of families of intermediate posi-
tion— skilled workmen, clerks, those in the lower ranks of the professions,
and so on — whose incomes were adequate, or nearly so, to cover costs,
and who needed but little direct assistance which would involve only
moderate and calculable sacrifices to the exchequer, although guarantees
might be required to make the structure of the necessary credit crisis-
proof. Slum clearance would then do the rest.
This policy was adopted by the Chamberlain Act (1923 to 1929;
amended by the Financial Provisions Act of 1924), which provided the
backbone of the house-building activity during our period. The Addison
Act (1919, restrictively amended 1921) had started the movement by
vesting the task of building houses with local authorities and guaranteeing
them against loss. The Wheatley Act (1924) again aimed chiefly in the
direction of building by local authorities. But the Chamberlain sub-
sidies really did the job. The effect of these measures, which eventually
burdened the exchequer with not more than about 13 million pounds a
year — a burden that was, moreover, compensated, at least in part, by the
increase in revenue and (relative) decrease in unemployment expenditure
which they at the time certainly entailed1 — was to raise the annual number
of units built above the annual average of the last prewar decade by
19242 and to a maximum of 273,230 units in 1926-1927, excluding houses
above 78 pounds or, in the metropolitan area, 105 pounds ratable value.
In all — counting in the tail end in the thirties, consisting of about 257,000
units built under the Wheatley Act — 1,177,863 subsidized houses were
erected at an estimated cost of 671 million pounds, local authorities being
responsible for 756,298 houses and 419 millions, private enterprise for the
rest. These subsidies tapered off from 1928 on and were, pro futuro,
abolished by the Housing Act of 1933, when this policy had been aban-
doned in favor of direct attack on the problem of the lowest strata by
means of slum clearing (Act of 1930). By then, however, conditions had
begun to shape for the expansion of unassisted building, which increased
in 1928, when subsidized construction fell off sharply — there was in the
total a net decrease, of course — continued to increase through 1931, and,
after falling but insignificantly in 1932, embarked upon its astounding
performance.
No problem arises, however, for our period. The, roughly, 71,000
houses built without assistance in 1928—1929 still compare with, roughly,
133,000 subsidized units and are, if anything, less than could have been
1 The theory of this is due, as the reader knows, to Mr. R. F. Kahn. In its bearings
upon our case, the theory of Secondary Employment is not invalidated by Professor
Neisser's criticism in Review of Economic Statistics for February 1936.
2 That prewar average was about 100,000. In 1922 over 106,000 houses were built
under the Addison Act, but this was due to an effort to profit from the act before the
restrictions of 1921 became effective.
1919-1929 753
expected from tlie fall in costs — building costs only — to little more than
one-third of the 1920 figure. Another factor, which in the thirties was
to play so great a role was, however, gathering power all the time. Build-
ing Societies, the history of which goes back to the last decades of the
eighteenth century and which, favored by public policy in many ways,
had made considerable headway since the eighties of the nineteenth,
increased their assets — mostly mortgages — which in 1918 were only about
5 per cent above what they had been in 1913, from 77.3 million pounds in
1919 to 312.7 in 1929. This increase, significant also for other purposes
than the one in hand, was as steady as it was great. Together with their
organization — ideally adapted to the financing of houses in the price
range from 400 to 500 pounds — and their privileges, this made them the
leaders in that line of business. Insurance companies were their chief
competitors. The success of both is an outstanding symptom of the
extent of the transfer of wealth that had occurred.
It is obvious, then, that during that period English residential build-
ing was primarily a function of subsidies and of expectation concerning
subsidies but not, as in the United States, of cyclical phases. It shaped
business1 situations rather than being shaped by them and thus "dis-
turbed" the course of cycles. Considering the quantitative importance
of this influence we shall hence expect what, from the standpoint of our
schema, would have to be called irregularities. As it happens, they are
not very great. Needless to say that, even if they were, they would
neither constitute a difficulty nor offer any basis for an objection to that
schema.
E. The "Industrial Revolution" of the Twenties. — These processes
were so entirely normal in the sense of conforming to expectation from
our model and so obviously repeated the history of preceding Kondratieff
downgrades that no war effects or other disturbances availed to obliter-
ate the fact, and that recalling a few familiar features suffices to
establish it.
First, we should not expect to find fundamentally new things, but rather
induced and completing development on lines chalked out before and
attended by strong increase in quantities, marked improvement in qual-
ities, "rationalization" all around, an indefinite number of individually
small innovations producing a wide variety of new specialties, the phe-
nomena which we have called conquest of new economic space. This is
what we find. The electrical, chemical, and automobile industries, which
together with their subsidiaries and all that directly and indirectly hinges
1 So did public construction, of course. For example, local authorities spent, alone
through their gas, water, and electric departments, about 17.5 million pounds in 1924.
Highways, harbors, wharves, docks, canals, sewers, embankments account for another 12.9
millions; roads and bridges, for 36.7.
754 BUSINESS CYCLES
upon them — the motorcar, for instance, is responsible for a great part of
the total of postwar construction: roads, garages, gasoline stations, sub-
urban residences — account for 90 per cent of the postwar changes in the
industrial organism and for most of the increase in real income. They
realized the possibilities created in the Kondratieff prosperity, continued
to push ahead from the bases laid before, and by so doing shaped things
into a Kondratieff recession. So did not only those subsidiaries, such as
oil and rubber, but also the minor, though still important, novelties, such
as steel alloys, aluminum, rayon, large-scale retailing, and the organiza-
tional and financial complement — persistence of the merger movement,
power finance and so on. There were exceptions, as there were in the
two previous Kondratieff downgrades, but none of them was quantita-
tively significant. The most important one was air transport on a com-
mercial scale, which may bear comparison with the role of railroads in the
thirties and of electricity in the eighties of the nineteenth century.
Second, we find all the general features which analysis and historical
observation have taught us to associate with Kondratieff downgrades.
This will become clear beyond doubt in our discussion of time series
which is to follow, but it should be clear independently of it, that those
features can be accounted for in terms of the system's absorption of and
reaction to the new quantities and new methods. We find prevalence of
unemployment1 that was, however many other circumstances may have
contributed to it, basically "technological." There was, though also
accentuated by other circumstances, that excess capacity2 which is
inseparable from the process of rapid reorganization of the industrial
apparatus and coexists with vigorous expansion of output. We observe
that desperate struggle of firms for outlets3 and against competition and
the sagging of prices incident to the insertion of new quantities and capaci-
ties, which understandably creates the picture of apparently permanent
1 On this and the two following points, see infra, Sec. F, I, a, 2 and 8.
2 As has been pointed out before, during a process of expansion and rearrangement of
this kind, normal utilization of capacity — if it be possible to speak of normals at all in
situations which, though moving toward normality, are yet far removed from any normality
— must always be considerably under 100 per cent, even apart from oligopolistic situations;
and this does not necessarily imply any waste, much less, at any rate, than would an
attempt "to maintain the profitable operation of moribund plants." See M. C. Rorty,
The Equation of Economic Balance, Harvard Business Review, April 1934. To repeat,
conditions in the production of raw materials reflect the same phenomenon in the form of
"overproduction.'* With varying degrees of severity, such conditions prevailed through-
out the period.
8 One symptom of this was the orgy of advertizing. Expenditure on newspaper adver-
tizing alone has for this country been estimated at a billion and a hah* for 1927, see Federal
Trade Commission, Resale Price Maintenance II, 1981. There is no possibility of reliable
comparison with prewar times, but the fact of steady increase of that item of expenditure
frpm 1922 on seems beyond doubt.
1919-1929 755
"overproduction" or "overinvestment" and the characteristic outcry
about people's inadequate power or willingness to spend. And, masked
and retarded by resistance to adjustments, the competing-down process
is clearly recognizable both within the relatively new and as between new
and old industries, railroads1 and coal being conspicuous instances of the
latter. All of which accounts for much of the social and business
atmosphere of the period, including its economic slogans.
Even in the comparatively "pure" case of the United States, "dis-
turbing factors" must be taken into account. In the cases of England
and Germany they are, of course, very much more important. Where
they work against our process they are, qualitatively at least, not difficult
to identify, and there is nothing objectionable in saying, for example,
that international restriction of rubber production partly stands instead
of that fall in rubber prices which would otherwise have ensued and that
its very existence proves the tendency which to some extent it succeeded
in paralyzing. Where, however, external factors tend to produce effects
similar to those that the prevailing phase of the cyclical process tends to
produce of itself, the difficulty sometimes spreads from measurement to
qualitative diagnosis. We shall meet several instances of this throughout
the rest of our way, the most controversial of them being the question
of the effects of recovery policy. Here a single example may suffice.
We have seen that in Germany factors extraneous to our process made
for what was called a consumers' prosperity. But consumers' pros-
perities are also part and parcel of Kondratieff recessions, and it can be
shown that some of the symptoms which that term is intended to cover,
would have emerged, perhaps not much less strikingly, in the absence of
those factors. Total effects are difficult to allocate.
1. English industry, as will be evident from the behavior of English
time series (see infra, Sec. F, I) displays the characteristic features of
the period much less markedly than American or German industry.
This, as has been explained in the preceding sections, must be understood
from the sociology of her case, from the way in which England's economic
position was affected by the war, and from developments in other coun-
tries that deprived her of economic elbowroom. The latter two factors
show in the fortunes of her export trade, which recovered slowly from the
drastic fall in 1921 — exports of manufactures and coal amounted to
1,220 million pounds in 1920 and to about half that sum in 1921 — and
never reached the physical volume of 1913 again. The latter increased
from 1925 to 1929 by a little more than 8 per cent (values in current
pounds fell), but this did not prevent a contribution of about 25 per cent
1 For England, cf. Mr. Gilbert Walker's exceedingly interesting study on the Economics
of Road and Rail Competition, Economic Journal, June 1988. A vast research program
lies in this direction.
756 BUSINESS CYCLES
to the unemployment of the latter year.1 Persistently favorable terms of
trade — due to the fall in the prices of raw materials, and in keeping with
cyclical expectation — must no doubt be set against this result, which
they, of course, helped to produce.
The task indicated by this development and by the presence of large
areas that were depressed for this and other reasons — reorientation of
production — was a specifically English one. It provided the scope for
the geographical migration — to the metropolitan area and the South —
and for an economic migration of resources into the diversified industries
catering for the middle and lower strata of domestic consumption, a
process which, since as usual the New did not grow out of the Old,
accounts for bright and dark spots alike. Within this picture, however,
and barring the influence of subsidized building — partly conditioned in
turn by the geographical migration, especially in the second half of the
period — the role of the outstanding leaders — electricity, chemistry, and
motorcars — is obvious. The motorcar industry, greatly helped by the
McKenna duties and road building, did much better than in Germany
and substantially contributed to England's new industrial structure in
the West Midlands and elsewhere. At first, prospects were most dis-
couraging. During the war the plants of motor (and cycle) concerns
had been much expanded, but on antiquated lines. They emerged in
a state of inefficiency and immediately lost domestic ground as well as a
great part of their export to America. Moreover, engineering firms whose
productive apparatus was capable of being adapted to the production of
motorcars then crowded into the market, which was, finally, upset by
the sales of the War Disposals Department. All this was an element in
the crisis of 1921. But a new development started in 1923 with the mass
production of cheap cars. From 1921, when but 40,000 private cars
and commercial vehicles were produced, output rose steadily, reaching
209,000 units in 1927.2 The number of insured persons engaged in pro-
duction and repair of motor vehicles, cycles, and aircraft rose from
192,000 in July 1923 to 233,000 in July 1927, while exports of complete
motor vehicles and especially of chassis far surpassed the 1913 numbers.
Some of the subsidiary industries and the aviation industry developed
still more vigorously. It is worth while mentioning that until 1925-1926
— when motor vehicles had increased by 200 per cent over 1919-1920 —
this went on amidst what, to fit their theories, some observers described
as "unrelieved gloom. " The nature of the commodity, more than the
quantitative importance of direct and indirect expenditure on it and
induced by it, disposes of this diagnosis.
1 See Mr. C. G. Clark's Statistical Studies Relating to the Present Economic Position
of Great Britain, Economic Journal, September 1931, for a discussion of the various problems
concerning exports.
8 The figures are as published by the Society of Motor Manufacturers and Traders.
1919-1929 757
The motorcar, aviation, and other expanding industries created new
demand for steel. But the iron and steel industry was nevertheless —
and in spite of the advantages it retained: good coke, low cost of trans-
portation for some of the world's highest grade ores — ailing all the time
and must be listed, along with coal and cotton, among the weak spots
of the organism. Expansion during the war, depressed conditions in
the shipbuilding industry, loss of foreign markets — directly and indirectly
almost three-quarters of the United Kingdom output used to be exported
— account for low prices and, in an old industry which follows rather than
leads, for the persistence of antiquated methods. But it is also true
that there was little of large-scale domestic investment in industrial plant
and machinery. Even the well-marked rise of steel consumption in 1927
left it, though nearly 30 per cent above the 1913 figure, distinctly low.1
We pass by chemistry — although it affords the best example of entre-
preneurial achievement on modern lines England has to show in that
period — and everything else, in order to note the fundamental role
electricity — again, barring building — played directly and by what it
induced. Development in this line, not a mere function of growth but
originating and propelling, carried output of the public supply system
of Great Britain from 2.5 billion kilowatt-hours to 11 billions at the end
of our period, when about 327 million pounds had been invested in the
generation, transmission, and distribution of current. This is less than
two-thirds of the cost per kilowatt installed in the United States. Tech-
nical improvement and larger generating stations (average capacity of
generating stations 12,000 kilowatts, as against 7,350 in the United States)
account for this as much as does the difference in price level and the lower
capital cost of steam-power plant. Average consumption of coal per
unit generated fell, much as in this country, from about 3.2 pounds in
1920 to 1.8 in 1930. English industry, which in 1914 had still resisted,
was definitively converted, and especially the many small and diversified
industries that were newly cropping up were electrified from the first.
But private enterprise supplied impulse and initiative only for the —
very considerable — development of the electrotechnical industries,
which, in the fashion of nineteenth-century capitalism, grew in a large
number of concerns until the great amalgamation into the Associated
Electrical Industries Ltd. (1929) put a more modern touch upon it.
In the power development, the initiating role belonged to the state, which
thus led in the two most important lines of advance that characterized
the period. The significance of this lies in the two facts, that entrepre-
neurial initiative, right from the beginning of the Kondratieff, signally
1 Home consumption of crude steel in 1918 was 2,181,000 tons quarterly average; in
1927, 2,695,000; in 1928, 2,417,000; in 1929, 2,662,000; a sharp fall thereafter until the
recovery from the second quarter of 1983. Cf. London and Cambridge Economic Service.
758 BUSINESS CYCLES
failed in an obvious and purely economic task, and that public agencies,
stepping into the breach, attacked it with perfect technological and
economic success, although, to be sure, with the entrepreneurial success in
America before their eyes. Only 27 per cent of the production of the pub-
lic supply system was privately owned at the end of our period. The
National Grid, recommended in 1925, was made an item of public policy
in 1926 (Electricity Supply Act).
Our sketch and the analysis underlying it can easily be verified by
a list of the industries that expanded — or, expanded more than others1 —
during the decade. Building, electricity — with constructional and
electrical engineering and many subsidiaries of both — and motor vehicles
(also aircraft) are there. Furniture follows in the wake of residential
construction, as does heating and ventilating apparatus. Rayon, of
course, was still an innovating industry. Presence of cement and mis-
cellaneous metals will not surprise us, nor will the presence of public
works. But then we find, testifying to the reorientation mentioned above
and characteristic of downgrade developments, a great expansion in the
miscellaneous trades and services, also in professional services and com-
modity distribution, in tobacco, food, drink, silk • • • while coal, other
engineering, cotton, iron and steel, shipbuilding, and railways supply the
complement.
A postwar boom in business and on the stock exchange, studded with
strikes, started in 1919 and lasted into the summer of 1920. While it was
going strong, banks hardly lived up to the idea some economists have
formed about their excessive bent for restriction.2 Nor, as has been
pointed out in a preceding section, was government policy conspicuous for
retrenchment. No further comment is required about the crash in 1921,
which was for England accentuated by its effects on her foreign trade and
which, to repeat, coincides with what according to our count would have
been the beginning of a Juglar depression. There was some recovery,
in business as well as on the stock exchange, in the fall, but the ground
gained was lost in the first half of 1922. Then building activity asserted
itself, the Kahn effect (secondary employment) of which presumably
helped to mitigate things in 1923. But there were other bright spots,
the wool industry being one of them, and the stock exchange was active.
The year 1924 displayed all the symptoms of a recovery weighed down
by unfavorable environmental conditions. Subsidized building lent
1The criterion is employment — no doubt an unreliable guide, the short-comings of
which account for the absence, excepting rayon, of the chemical industry, but which may
pass muster for the purpose in hand.
2 The severity of the subsequent breakdown illustrates well the consequences of "liberal"
credit policies. It should be particularly noticed that the Bank of England gave an impulse
in this direction by repaying Special Deposits. See S. E. Harris, Monetary Problems of the
British Empire, 1931, Chap. III.
1919-1929 759
support, however. Imports increased, prices rose somewhat, unemploy-
ment decreased. Even the shipbuilding and cotton situations were
relieved. 1925 brought expansion in many of the newer industries, but
what might have been the completion of recovery was greatly interfered
with by labor difficulties (coal, shipping, wool) and the numerous partic-
ular depressions in individual industries, i.e., depressions due to causes
peculiar to them.
The clearly abnormal situation of 1926 makes the year difficult to
class. A struggle that in any other country might have spelled revolution
paralyzed everything. The difficulty extends to 1927, because the
repairing of the omissions of 1926 and the fact that 1927 marks the peak
of subsidized building would in themselves suffice to explain the brighter
colors of that year. However, we do not rely on aggregates and indices
only, but also and primarily on what happened in the industrial organism.
There many of the new things mentioned above now began to gather
momentum. It does not seem unreasonable, therefore, to speak of a
new — the fourth — Juglar, since this is not more than our way of expressing
those very facts. And it will, perhaps, also seem acceptable to date it
for the purpose of counting from 1926, seeing that in the absence of a
clearly abnormal event the prosperity phase would presumably, condi-
tions in 1925 having been what they were, have set in earlier. There is
no reason, however, why we should insist on this.
2. In Germany, the term Rationalization was used more commonly
than it was anywhere else, in order to describe the industrial processes of
the postinflation period. It not only expressed what amounted to a
conscious national effort sponsored by all classes — though partly from
different motives — and encouraged by the federal government,1 but it also
expresses the gist of what we mean by downgrade developments : exploita-
tion to the utmost, partly under duress, of existing possibilities of tech-
nological and organizational innovations on lines and principles estab-
lished before but steadily improved in the process; revision of the whole
structure of industry in quest of increased efficiency; systematic struggle
with each item of the list of costs — all of which is exemplified to perfection
by the postwar history of all branches of German industry. Few things
1 M ore burocratico, this encouragement found expression in a new governmental board
called the Reichskuratorium fair Wirtschaftlichkeit, the first achievement of which con-
sisted, characteristically enough, in defining the concept of rationalization. It is, however,
more important to note that the spokesmen of the trade unions were by no means hostile.
They sometimes maneuvered themselves into somewhat difficult positions by arguing that
high wages would make themselves possible by enforcing rationalization and otherwise
espousing the cause of technological improvement, while, of course, allegiance was at the
same time due to the doctrine that such improvement necessarily injures the interests of
labor. Capitalism was condemned for innovating too slowly and too quickly at the same
time. But no actual resistance was offered on the latter ground.
760 BUSINESS CYCLES
were fundamentally new, the most important items being1 the production
on a large scale of synthetic nitrogen (the Leunawerke of the Dye Trust
were, however, erected in 1916 as a war measure; the Haber-Bosch inven-
tion dates from 1913), the production of aluminum, which in Germany
also dated from the war, and, on a commercial scale at least, the radio
and the aeroplane. Nevertheless, there was as complete a transforma-
tion of the economic organism as there was in England after the Napo-
leonic wars.
For purposes other than ours it would be necessary to dwell on the
great and increasing role played by the industrial activity of the Reich,
the states, and the municipalities. As far as the Reich was concerned, it
had acquired most of its business property during the war and with no
intention beyond providing for war requirements, although both the logic
of the situation thus created and the spirit of the times made it easier to
expand than to liquidate it. But the states had inherited important
interests from their monarchic predecessors and added to them as a matter
of principle, while the municipalities simply went further along the path
of " municipal socialism." For 1925 total turnover of all public enterprise
was estimated by Dr. Marschak at 10 billion marks2 and total value of
all publicly owned property by Professor J. Hirsch — conservatively —
at 52 billions, roughly one-fifth of the national total. However, 26 bil-
lions are accounted for by the federal railways — which were merely
transferred, as a consequence of postwar arrangements, from the states
that previously owned them to a corporation owned by the Reich — and
1 billion by the federal post office. The other concerns mainly or wholly
owned by the Reich were combined into a holding company (Vereinigte
Industrie-Unternehmungen A.G.) and included electrical and electro-
technical, aluminum — the bulk of the German aluminum production was
controlled by the Vereinigte Aluminium Werke wholly owned by the
Reich — nitrogen, iron- and steelworks, and mining. The Reich also
founded a commercial bank of first rank, the Reichskredit-Gesellschaft.3
Similarly, the states strengthened their foothold, especially in the
fields of electricity and mining, but also in others. The municipalities
built and operated additional gas, water, electrical works, street railways,
slaughterhouses, and so on, but did not materially go beyond utilities and
residential building. The interesting thing about those industrial
1 We do not list the hydrogenation of coal (or the staple fiber) because it did not, during
the twenties, play any significant role, though the technological bases were being laid for
the developments of the thirties. This was done within the Dye Trust (Bergius process).
2 See Wirtschaftsdemokratie, ed. by the Allgemeine Deutsche Gewerkschaftsbund, 3d
ed., 1929, Appendix I.
3 The states and other public bodies did not lag behind. It is interesting to note that
total assets of banks controlled by public bodies were, in 1927, 12.7 billion marks as against
the 16.1 billions of all other banks.
1919-1929 761
properties of the Reich and the states is the thoroughly businesslike
manner in which they were managed. The public authority was a stock-
holder, sometimes but not necessarily the only or the controlling one, and
interfered but little. The managements enjoyed not much less inde-
pendence than they would have in any ordinary case, and they behaved
as managements ordinarily do — they took pride in technological perfec-
tion, good profits, and sizable reserves. The whole arrangement seemed
devised in order to put competent businessmen in charge and to keep
politics at bay. As a possible solution of the problem of industrial
leadership in a socializing state — a problem that loomed large in the
discussions of the Commission on Socialization in 1919 — it seems to merit
attention.
For our purpose, however, we need not take account of this element in
the economic life of Germany, except in the case of the production of
electrical power which, vigorously developing even during the period of
inflation, reached 20.3 billion kilowatt-hours in 1925, 25.1 in 1927, and
30.7 in 1929. Of this total, from 2 to 4 billions were produced in hydro-
electric plants. Lignite1 was used in the production of about one-third
of it. About 70 per cent of the output of public supply stations — not of
the totals mentioned — was controlled by either the Reich or the states or
by municipalities, although the relative share of the latter fell, as com-
pared with prewar times, owing to the improvements in long-distance
transmission, which was unfavorable to local sources of moderate size.
The federal governments power concerns and participations in power
concerns, of which the A.G. fiir Elektrizitatswirtschaft and the Rheinisch-
Westfaelisches Elektrizitatswerk were the most important, were paralleled
by the Preussische Elektrizitats A.G., which combined the power inter-
ests of the Prussian state, and by the works owned by Saxony, Bavaria,
Baden, and Thuringia. This alone, and apart from the municipal sta-
tions, practically amounted to public control of generation and transmis-
sion and, of course, foreshadowed progress toward a perfectly coordinated
system of interconnecting superpower stations. The work of electrifica-
tion was, however, not nearly completed even in industry, while agricul-
ture in 1925 consumed only 0.4 billion kilowatt-hours of electrical energy,
and railways were, even in 1928, electrified only to 2.4 per cent of track-
age. The electrification of the household was hardly begun. As far as
the writer is able to make out, total investment was, 1925 to 1929, of the
order of magnitude of 2 billion marks.2
1 Brickett production for domestic fueling, the use of lignite as a chemical raw mate-
rial, and its use for the production of electric current account for its spectacular career.
Mechanization of lignite mining, of course, helped. Horsepower installed rose from 55,000
in 1895 to 209,000 in 1907 and 766,000 in 1925; output, which increased even during the
war, from 87.2 million metric tons in 1913 to 139.7 in 1925 and 165.6 in 1928.
2 Assuming that investment per kilowatt installed was about 1,000 marks.
762 BUSINESS CYCLES
Value of output of the electrotechnical industry made a big stride
between 1925, when it was 2.1, and 1927, when it was 2.7 billion marks.1
These are very hazardous estimates, but they are fully borne out by the
official and more reliable figures of value of exports. They rose only by
little more than 20 per cent during those years3 — to 441.2 millions or
26.7 per cent of the total for all countries, which compares with the
417.5 million marks and a 25.2 per cent share of the United States. But
if that was possible in the face of the existing barriers and the fact that
export in this line requires a great deal of capital, it is safe to assume that
domestic sales increased much more strongly. Innovation was effec-
tively corraled, however. There were new concerns, but the two that
had been leading before the war, Siemens and the General Electric,
retained that position.
In 1913 the machinery industry — excluding electrotechnical produc-
tion and boilers but including locomotives — worked at practically 100
per cent capacity, producing products of a value of 2.7 billion marks
(postwar territory) to which export contributed 0.74 billions. In 1925
value of production at prewar prices has been estimated at about 1.9 bil-
lions while capacity was for 3.36. 3 Progress during 1925-1929 in quantity
and quality of output was very great, value of products at current prices
rising by 38 per cent, considerably more than anywhere else, and exports,
which rose to about 1 billion marks by 1928, recovering part of their pre-
war position.4 While in many respects this was a great entrepreneurial
achievement — or rather, the result of a long series of entrepreneurial
achievements within a large number of medium-sized firms — which con-
tributed substantially to the economic processes of the time, and while,
though of smaller quantitative importance, the industries producing
optical, medical, and other instruments, photographic apparatus, and
so on — typewriters have been included in machinery — kept well in step,
the automobile industry stagnated, notwithstanding prohibition of
1 Comparison would be greatly to the disadvantage of the United States, where that
value rose but moderately during the same time. This may, however, be as much due to
the faultiness of estimates as to the fact that in this country the big stride had been made
before 1925.
2 However, 1928 brought a great increase, although Germany's share in the world's
trade still remained far below the 50 per cent of 1913.
3 Cf. the Denkschrift tiber die Maschinenindustrie der Welt, October 1926, prepared
for the League's International Economic Conference by the Association of German Pro-
ducers of Machines (Verein Deutscher Maschinenbau-Anstalten). With due respect for
this excellent memorandum, to which the writer is much indebted for a better grasp of the
situation of that industry in 1925, exception must be taken to the two methods suggested,
p. 17, for arriving at an estimate of capacity. For the estimates actually made it may be
urged, however, that the officers of the Association were likely to have a pretty good
impression.
4 United States exports were 1,688 million marks in 1928.
1919-1929 763
imports followed by heavy protection, until about 1927, not only because
of obvious postwar difficulties but also because of a temporary inability
to modernize itself which is not completely explained by those difficulties.
The old firms, which before the war had been among the international
leaders, lingered on or died, and only inefficient new ones were added for a
time until the belated advent of the small cheap car that was sufficiently
economical as regards taxes and gasoline. Domestic progress was also
quickened by General Motors acquiring and Ford erecting plants in
Germany, prices eventually fell to a level on which the homemade car
had a chance to sell — 1928, the year of maximum importation, is the
first year in which this can be said to have been the case — and the weak-
est firms dropped out. We therefore note that both investment and sales
began to be significant in the second half of our period; but, stated in
absolute figures, success remained small. To the 20,000 units which,
including trucks and buses, were produced in 1913 corresponded not
quite 63,000 in 1925 and about 138,000 in 1928. Motor cycles did better.
Mechanization (e.g., of rolling), concentration of production in bigger
units optimally located (e.g., Hamborn), varied progress in electrometal-
lurgy, standardization of products, and improvements in the use of heat
and power are the familiar features of typically induced developments in
mining and the heavy industries. We will confine ourselves to the
organizational aspect. It has been pointed out in Chap. VII that infla-
tion gave a powerful impetus to the merger movement, which already
had gone very far during the Kondratieff prosperity. The losses of terri-
tory, often cutting through the domains of concerns and upsetting
established relations between materials or stages of production, also
necessitated reorientation. Many unshapely and unmanageable mon-
sters resulted, which were unable to live as soon as the contours of reality
emerged from the fog. The breakdown of the Stinnes concern, the dis-
solution of the Siemens-Rheinelbe-Union are but conspicuous instances
of a process of liquidation which was referred to as the "crisis of concerns"1
and which, even where it did not lead to failure, put an end to many vertical
and horizontal combinations, participations, understandings, and so on.
Even such concerns as Krupp, Stumm, and Rombach, had to retrace
steps and to take losses. The ordinary and the international cartel and
regulation by public authority (Zwangssyndizierung) were the remedies
resorted to. As an example for the latter kind of rationalization we will
mention the potash industry.2
In this case it had already started during the war, when the emergence
of new works and the sinking of new pits was barred by government
1 The German word Konzern means combination rather than firm.
2 Of. the report of the Survey on Economic Conditions (Wirtschaftsenquete) on that
industry. This report also appeared as a book, 1929,
764 BUSINESS CYCLES
decree (1916). Three further decrees (1919, 1920, and 1921) put a heavy
premium on the closing of mines, which was also facilitated by the merger
movement in progress. While the number of works enjoying quota
increased from 1921 to 1928, when there were 229 of them, the number of
works in operation in the same period decreased to 60 which, greatly
improving methods and the quality of the product — also 82 per cent of
the product was by 1928 no longer sold in the crude state but processed
within the industry, by-products gaining steadily in importance — and
reducing cost, increased their output more than threefold in the course
of five years.1 The mergers left six independent concerns in business,
of which three produced 80 per cent of the national total. Sales were
centralized in the potash syndicate, which in 1925 negotiated an English
loan and, according to a complicated schedule, granted rebates to the
various (five) classes of domestic buyers but did not fix the prices. These
were fixed by a public authority, the federal potash council (Reichs-
kalirat), and subject to the approval of the Ministry of Economic
Affairs — needless to ask whether or why they were rigid. In inter-
preting this case, which is of interest far beyond its relevance to our
subject,2 the reader should, however, bear in mind the fact that capital
invested had, before the war, been estimated — very unreliably to be sure
— at 1.4 billion marks of prewar purchasing power, while in the Survey
of Economic Conditions it was for the end of our period estimated at from
600 to 700 million marks of postwar purchasing power — a loss which no
doubt comes under the headings of malinvestment and waste of
competition.
Several other cases of more or less the same type — coal3 and brown-
coal mining for instance — could be cited. We will, however, confine
ourselves to the case of iron and steel. It could also serve as an example
of international regulation which suggested itself owing to the great
expansion in capacity coupled with both improvement and more eco-
nomical use of the product all over the world. Mere quantities, very
clearly indicating the upswing that lasted from 1925 to 1928, increased
during those years by 15 per cent (iron) and 20 per cent (steel), in Europe
alone by respectively 25 and 30 per cent. German production of steel
surpassed its prewar level of not quite 12 million tons (1913, postwar
1 The output of the industry rose to about 10 per cent above prewar level in 1925 and
further increased in 1928. Costs of power and heat in terms of lignite fell to less than half.
Workmen employed fell, 1923 to 1928, from 40,000 to 19,000. Wage rates rose over 60 per
cent from 1924 to 1928, so that wage bill and labor cost per unit of product fell.
2 The measures taken and the way in which they worked out should in particular be
interesting to American students of planning of the NRA type. The problem dealt with
by those measures was exactly analogous to the one facing American policy in the time of
the NRA.
3 Coal Act, setting up the Federal Coal Council (Reichskohlenrat) of Apr. 24, 1919.
1919-1929 765
territory) in 1925, and was over 16 million tons in 1927,1 though the
production of foundries in 1927 was only at the 1913 figure (exactly).
The production of rolling mills (semifinished and finished products)
was in the same year far above the 1913 figure (by about 60 and 30 per
cent respectively2). The Berlin Institute's index of iron prices which in
1925 stood at 125 per cent of 1913, fell in 1926 to 112 per cent and then
rose, with a setback in 1928, through 1930.
For us the interesting point is that that development occurred in the
course of an "organizational rationalization" which was effected exclu-
sively by entrepreneurial effort from within the industry and is suggestive
of earlier American examples. The new Vereinigte Stahlwerke, which
attained corporate existence in 1926, were a unit of control that aimed at
concentration and specialization of production in optimally located plant.
From 1926 to 1933 (when in the wake of the crisis another reorientation
and reorganization took place) pits were reduced from 48 to 25, iron-
works from 140 to 66, steel foundries from 20 to 8, rolling mills from
17 to 10. Very considerable investment was required to achieve this, and
almost immediately after their foundation the Stahlwerke incurred a
debt of over 500 million marks, a little more than half of which was spent
on the erection of new plant and the improvement of existing plant.
Their quota in the syndicate was less than 40 per cent, however. Not
only such firms as Krupp, Mannesmann, and Hoesch retained their
independence, forming other alliances and expanding on their own, but
also another group was formed under the leadership of the Rheinelbe-
Union, which embarked on an extensive investment program of similar
type3 and soon floated an issue of 800 million marks. The works — those
belonging to different combinations as well as independent ones — which
produced steel specialties, in 1927 formed an organization called Deutsche
Edelstahlwerke. Every one of these steps was accompanied by induced
innovation of the technological type, and all of them created an almost
completely new industrial organism. That in the subsequent crisis
things should have presented a picture which it was as easy as it was
superficial to describe in terms of excess capacity and malinvestment will
not surprise us.
The quantitatively most important event in the chemical industry,
which carried everything before it and solved the problem created by
the loss of patents after the war, has been mentioned already. Mark
volume of chemical production rose by 33 per cent from 1924 to 1928,
1 At the same time the number of workmen employed by steelworks was lower by about
16 per cent.
2 It is for our purposes important to note that there was a sharp setback in 1928.
8 The above refers to the western district. The heavy industry of Upjter Silesia com-
bined independently, and the steelworks of central Germany (Mitteldeutsche Stahlwerke)
formed yet another group in which the Vereinigte Stahlwerke participated to 50 per cent.
766 BUSINESS CYCLES
exports went over 1 billion by 1926 — first indications of the unfathomable
possibilities of synthetic materials. No other industry displays so
clearly an ineluctable necessity of largest scale enterprise — the J. G.
Farben (Dye Trust) produces about 100 per cent of total German output
of dyes, about 85 per cent of the output of synthetic nitrogen, and about
90 per cent of the output of sulphuric acid, and the Imperial Chemical
Industries Ltd. in England, the fitablissements Kuhlmann in France,
Montecatini in Italy and, to a lesser extent, Du Pont de Nemours in this
country hold comparable positions — as well as the necessity, for profits
and even for survival, of incessant innovation, as does this premier indus-
try of the future. Into its problems we cannot enter any more than into
the problems of the rayon industry, which during our period made its
decisive stride, consumption about doubling from 1925 to 1927. This
compares with all but stagnation in the cotton, wool, linen, and jute
industries. The marked upswing in 1927, which was well sustained
afterward, should be mentioned, however. And so should the concen-
tration which went on and reduced, for example, in the cotton industry
the 21,600 units that existed in 1907 to 8,000 by 1925. The fact that the
limits set to these fragments of a sketch do not allow us to enter into the
history of other industries is particularly regrettable, because it makes it
impossible to describe those items which would reflect the consumers'
prosperity of that time and that large medley of individually small
innovations — many of them little more than successful insertions of some
new article or brand by means of dashing advertizing campaigns — which
in Germany, as elsewhere, contributed so much that is most character-
istic of its processes.
Dr. Clausing,1 applying Professor Spiethoff's model, counts two
cycles, the first from a trough in November 1923 to another trough in
January 1926, the second from an incipient recovery in February 1926
to a depression (Niedergang) from the second half of 1929 to 1932. From
our standpoint there is very little to add to or to criticize in this. There
is no doubt that 1927 was a year of prosperity also in our sense. As the
reader will easily verify by reference to what has been said above, this
prosperity links up with industrial innovation which displays all the
characteristics we associate with Kondratieff downgrades. We shall see
later that time-series evidence would support that: unemployment
decreased strongly; price level rose; liquidations, receiverships, and
1 Gustav Clausing, Die Wirtschaftlichen Wechsellagen von 1919 bis 1932, 1983. C. T.
Schmidt, German Business Cycles. 1924-1933, published by the National Bureau of Eco-
nomic Research, 1934, also observed two cycles, the one from a trough in December 1923
to a trough in March 1926, culminating about March 1925, the other from approximately
April 1926 to the late summer of 1932, the crest being difficult to establish (pp. 169 and
170). He notes that after the beginning of 1927 a "growing international similarity is
present." (Page 248.)
1919-1929 767
bankruptcies fell to their postwar minima; foundations of new firms
increased and so did corporate earnings. It was clearly the familiar pic-
ture of what we are in the habit of calling a Juglar prosperity. Although
we have decided (Chap. VII) not to go on counting Juglars in the case
of Germany, it is worth while to note that it would not be impossible to
do so. Unlike this country, and much more than England, Germany
experienced in 1928 what most people will agree to call recession, but the
symptoms do not quite answer to our conception of that phase and are
in some respects contradictory. The Berlin Institute's index of produc-
tion and employment fell, while other factors, e.g., money in circulation,
cost of living, and wholesale prices rose a little. There was, however, a
rally at the end of the year which lasted through the first four months of
1929, when the system, slowly at first, began to slide off into depression.
And the case for expecting reaction to the preceding industrial revolution
remains strong.
That prosperity did not start with 1927, but was in full swing at the
beginning of that year. In terms of industrial innovation and investment
as well as of many aggregative indices, such as stock prices, imports of
raw materials, orders received, total output, pig-iron production, employ-
ment, the second half of 1926 was a time of prosperity in our sense and
must be considered as the beginning of the upswing discussed in the
preceding paragraph. What went before, however, should not be
interpreted as a cycle in the sense of either our own or any other analytical
schema. The year 1924 starts — from a crisis and heavy unemployment
in the winter 1923-1924 — with an upswing which means little more than
that business, having fallen from the clouds of inflation, was trying to
find its feet. It was cut short in the middle of that year by the effects
of taxation and of a pull of the Reichsbank's curb. Prospects of foreign
credits — and the first foreign credits actually given — together with sham
profits due to low evaluation of assets in terms of the new mark then
account for a spurt toward normal volumes, which lasted from about
July 1924 to about February 1925. A severe relapse — in the fourth
quarter, particularly severe in employment — followed upon this, which
was due to a variety of circumstances peculiar to the situation. Hence,
it would not do to identify that up and down as a Kitchin.
3. In the United States conformity to expectation during that period
is, as stated above, so obvious as to make it almost superfluous to prove
it, a fact the value of which is enhanced by the — relatively speaking —
small importance of external disturbances in our sense. That the events
in the fields of electricity, motorcars, and chemistry do not, in our ter-
minology, constitute fundamentally new but induced and completing
developments, which proceeded from bases laid in, roughly, the two
prewar decades, needs additional emphasis as little as does the fact that
768 BUSINESS CYCLES
it was those developments that "carried" the economic processes of the
period. We may, however, note the substantive novelty of aviation as a
commercial success — 1925 may serve as a date — which was perhaps the
most important exception. The interesting thing about this industry is
that it developed on its own and not, as might have been expected from
standpoints other than ours, as an appendage1 to an older, say the auto-
mobile, industry in spite of the similarity its problems bear to those of
the latter. Exactly as the telephone industry was not built up by the
telegraph industry and has shown no tendency to be dominated by it,
and as the rise of the automobile industry owed but little to the carriage
and bicycle industries — or, we may add, to the firms that previously
produced the Otto motors — or as the moving picture industry, which we
might also list among the genuine innovations of the period, did its own
pioneering and was not the work, technically, financially, or commercially,
of the theater interests, so aviation supplies another instance in verifica-
tion of the hypothesis of New Firms and New Men (Chap. Ill) arising
independently of the Old Firms and laying themselves alongside of them.2
The same often holds true of new specialties within each great line of
advance, as within the field of electrical industries, partly at least, in
the cases of the radio and of the refrigerator.
a. Power production increased from 38.9 billion kilowatt-hours in
1919 to over 97 in 1929,3 only 1921 marking a relapse of about 8 per
cent. Roughly 95 per cent of this was produced by privately financed
enterprise and over half of it by the General Electric, Insull, Morgan,
Mellon, Byllesby and Doherty groups and a dozen corporations jointly
controlled by these. Although the more remote effects of this develop-
ment on industrial activity in general were much more important factors
1 This has been stressed so well by Professor M. W. Watkins, The Aviation Industry,
Journal of Political Economy, February 1931, that we cannot do better than quote from
that paper. The rest of the above paragraph is almost bodily taken from pp. 67 and 68.
2 Professor Watkins' comments on the phenomenon should be quoted: "The explana-
tion seems to be that the managers and directors of older industries, once they have suc-
ceeded in establishing as an economic 'going concern' the special branch of industry with
which they are primarily identified, lose their adventurous inclinations. They tend to
become skeptical of new processes and new products. They lose the Vision ' of industrial
pioneers. They become absorbed in the complicated routine of their own affairs and the
ever recurring problems of adjustment and adaptation of which no field of business enter-
prise is free. In these circumstances, it is only the far-sighted, uneasy ', venturesome indi-
viduals here and there who are ready to 'cut loose from' a secure position and assured income,
and who have the gift of imparting their enthusiasm to other restless individuals (technicians,
salesmen, laborers) and to still others, with private capital, [the present writer's italics] who
are willing to take great risks for the chance of great gains — it is only, in a word, adventurers
who found new industries. The aviation industry has been no exception • • • "
3 These figures are as given by the United States Geological Survey and differ considera-
bly from the figures of both the census and the National Electric Light Association.
1919-1929 769
in the cyclical variations of the period than the immediate effects of the
investment in power plant, transmission lines, and distribution, we may
yet note that from 1917 to 1927 balance-sheet values of power plants
increased from about 3 to about 9.4 billions1 and that more than 1.5
billions of electrical stock and bonds was issued in the yearly average
between 1924 and 1930 — the maximum of 2,150 millions occurred in
1927 — of which perhaps something less than two-thirds was spent on new
construction and extension.2 Gross earnings of the electric light and
power industry reached 2.1 billion dollars by 1929, 3 when household
consumption was responsible for 604 millions, industrial and commercial
for about 1.2 billion, street lighting and traction for the rest.
Prices, of course, differed widely, not only locally but also as between
customers: in 1929 the leather industry, for instance, paid $28 for 1,000
kilowatt-hours and the chemical industry 5.9, 12.7 dollars being the
average for that year as given by the census. In the average, however,
they fell. The national average price of current used in households is a
no less doubtful matter. The semiofficial figures are per kilowatt-hour:
16.2 cents around the turn of the century, about 9 cents for 1912, roughly
7.5 at the beginning of our period, during which it slowly but steadily
fell to 6.3 in 1929, or about 3.8 cents in terms of the prewar purchasing
power of the wage earners' dollar.4 This behavior of prices is accounted
for, on the one hand, not only by the actual or potential competition of
industrial — as distinguished from "public" — stations, but also by "com-
modity competition" — gas, nonelectrical motors — and the necessity of
building up new demand: the electrification of the household and of the
farm in particular was to a large extent a question of price. On the
other hand, the growth of units of control and the establishment of local
1 Census data; we might also compare the 1002 figure of investment in that sense —
about 0.5 billions — to the 1932 figure — nearly 13 billions, about $384 per kilowatt installed.
But not much confidence can be placed in any of those data. Variations in the significance
of the monetary unit apart, the figures of different census are not strictly comparable,
because they do not include exactly the same things. The above comparison can at best
give an idea about orders of magnitude.
2 About two billions of the total was raised between 1922 and 1932 from customers.
The — on the whole — rising standing of the industry drew new types of investors toward
its mortgage securities, as life insurance companies and also — compare, for example, the
New York State Act of 1927 — savings banks. In its then dimensions this was a postwar
development. Including the gas industry, total investment by all classes of investors is
said to have reached the figure of 18 billions by 1932, including what may be termed th<>
distress issues of that year.
3 The development is interesting: 1902, 85.7 millions; 1907, 175.6; 1912, 302; 1917, 521;
1922, 1,072; 1929, 2,107. Cf. Wall Street Journal, June 28, 1930.
4 According to the cost-of-living index of the National Industrial Conference Board.
The Department of Labor's prices of current are somewhat higher, but their tendency is
the same.
770 BUSINESS CYCLES
and sectional monopolies facilitated discrimination and went far toward
eliminating price competition between those units, while their struggles
were transferred to the financial sphere. That explains why the weighted
average of prices did not, within our period, fall correspondingly to the
increase in efficiency of production, and this again was why most oper-
ating companies were in a position to improve their financial status
considerably1 and to weather the subsequent storm comparatively well.
The competing-down process and its contribution to the general picture
of the period, but especially to the subsequent Great Depression, took
under the circumstances a form which was in many respects peculiar.
It asserted itself mainly through shifts in industrial location — electrical
development materially helping, for instance, in the industrialization of
the South — and much less directly as, for instance, in the effect on coal.
But no difficulty arises in elaborating this aspect.
Technological advance was much on the same lines as in Europe.
Water-power development played, of course, a great role: from 1924 to
1928 it progressed at a greater rate than the capacity of steam plants,
reaching an output of 29 billion kilowatt -hours by 1930 though at the
end of the period steam began to gain ground relatively. The use of
fuel oil and gas was an American peculiarity.2 Otherwise we observe
the general tendencies toward larger capacity of stations — the number of
plants fell by one-third between 1922 and 1929 — and superpower zones.
Since in extending electrical enterprise to foreign countries capital counts
for almost everything, the success of American groups, especially in
South America, is easy to understand (American and Foreign Power Co.).
About one billion went to South America, Europe, Asia, and into what
presently turned out to be so many traps.
Considering the technological nature of much that was done, mergers,
partly also aiming at the control of gas concerns, were the unavoidable
concomitant of this development. The financial instrument of the hold-
ing company lying ready at hand therefore experienced a new vogue of
unprecedented dimensions. Power finance definitively passed out of the
hands of the manufacturing industry and coordination resulted from a
struggle within the power-producing sphere, in which the groups men-
tioned above emerged or conquered. Since this struggle involved com-
1 The general question of raising funds for consolidation and investment — in part — by
"taxing" consumers tw. borrowing is conveniently discussed by means of the model of a
socialist state faced by the analogous problem. It cannot be disposed of by an argument
on the lines of Marshall's concept of consumers' rent because this covers at best the sta-
tionary case. All that matters for us, however, is the fact of that price stability and its
possible effects on the cyclical process under the three headings of "rigidity,'* profits, and
financial status.
2 The use of natural gas, itself a major feature of the period, increased in the decade
after 1919 from 21.4 billion cubic feet to 77.
1919-1929 771
petitive bidding for strategic positions, such geographical and commercial
rationalization as was achieved was accompanied by the growth of a
huge structure of debt — closed mortgages, open-end mortgages, insured
debentures — and share capital, which was out of proportion with the
effects of that rationalization, and not only provided food for purely
financial maneuvers and speculative excesses of a type suggestive of the
railroad age, but also jeopardized the banking system, since power securi-
ties loomed large in its collateral and since many leading banks, among
them the National City, the Chase National, the Bankers' Trust, the
Guaranty Trust, associated their fortunes directly with power enterprise
and in fact functioned in some cases as the agents of ultimate centraliza-
tion. Without going further into this well-known matter, we will note
that the great boom in power finance — and real investment — belongs to
the second half of our decade. It was a feature of the fourth Juglar and
clearly basic to its prosperity phase. In fact, building construction,
power development — together with developments in other branches of
the utility field which we cannot stay to discuss but which also fit into
our general idea of the processes of a Kondratieff downgrade1 — would in
themselves suffice to account for the behavior of aggregative time series
during the period.
The major instances of the propelling and dislocating effects of power
developments are obvious, and description of the sum total of all the
minor ones is impossible within this sketch. But it should be emphasized
in view of popular dirges about lack of investment opportunity that the
work of electrification — as much of it even as is technologically and
commercially possible at the moment or in immediate prospect — is not
nearly completed. There is enough investment opportunity from this
source alone for many a cycle to come. Even industry is as yet but
imperfectly electrified — perhaps to something like 75 per cent — and so
are households, while but a beginning has been made in the electrifica-
tion of farms and of transportation. Only the telephone and electric
lighting can reasonably be said to have exhausted, ex visu of present
technology, the bulk of their possibilities,2 although the automatic tele-
1 Utility developments form part of the picture which we expect a Kondratieff down-
grade to reveal, because they are to a large extent a function of real income and its rate
of change. Accordingly, we find expansion — induced expansion — in the utility field in
the two last decades of the first Kondratieff as well as in the downgrade of the second
(eighties and early nineties). We find the same phenomenon in the present instance.
2 A criterion of whether or not an industry is past its first spurt is in some cases afforded
by the presence or absence of setbacks in its expansion. As far as the writer is able to
make out, the number of telephones installed increased without any break, depression or
no depression, from 1876, when there was no commercial installation, to 1980 inclusive,
after which year there was a fall. Recovery set in in 1934, but the precrisis figure was not
nearly attained in 1935. However, within our period the number of telephones installed
772 BUSINESS CYCLES
phone — installation was zero in 1892 and only 1.7 per cent of the total of
telephones installed in 1919, but nearly 32 per cent in 1930 — which must
be listed among the innovations of the period under discussion, affords
a good illustration, if one be needed, for the fact that even perfect satura-
tion of existing demand need not call a halt of "progress."
Production of electrical equipment had, ever since 1915, increased at
a greater rate than production of power and continued to do so until
1929. Its value was about 1 billion dollars in 1919 and nearly 2.5 billions
in 1929. 1 Examples of new industries — and the "diversifying" effect of
power production — abound. We will merely note that the spectacular
expansion of the radio and the refrigerator industries dates from 1926.
The quarter of a million socket radios then in use increased to over
7 millions, the 315,000 refrigerators to 1,680,000 in 1929.2 Though
typical instances of downgrade developments, these were practically
new industries with histories of their own. But they were not so inde-
pendent of the older concerns in the industry of electric manufacture as,
say, aviation is of the automobile industry. Generally speaking, these
older concerns maintained their position well, and proved in this as in
other countries successful shells of incessant innovation, especially in
the heavy-current field (General Electric, Westinghouse) . Dollar volume
of output in electric manufacturing increased about sevenfold between
1914 and 1929, arid about 26 times from 1899, the census year nearest
to the beginning of the Kondratieff, to 1929.3
b. The automobile industry led in every upswing and out of every
downswing throughout the period, in fact beyond it, and continued in
the Kondratieff recession to qualify as well for the role of standard
example for the processes embodied in our model as it had done in the
upswing. Employment in motor-vehicle factories, not including pro-
duction of parts, tires, and bodies, increased from about 253,000 in
1922 to 427,500 in 1929, the corresponding wage bill from about 396 to
about 775.5 million dollars. Passenger-car registration as of Dec. 31
increased without any break from the beginning of the series (1895:4)
rose from about 12.7 millions in 1919 to over 20 millions in 1929. This increase was
sufficient to raise the number of telephone operators by about 80 per cent, in spite of the
labor-saving effects of the automatic telephone and some " taylorizing."
1 Census figures.
2 Figures of the Edison Institute. The number of socket radios continued to increase
throughout the depression to nearly 20 millions in 1935, an example for those initial spurts
which are impervious to depressions. As the reader will remember, such behavior is, if
anything, normal from the standpoint of our analysis, though in practice it is not the
general rule. Similarly, the number of refrigerators in occupied homes kept on increasing
without a break and reached the figure of 7.25 millions in 1935.
8 The number of workmen employed rose from 42,000 to 329,000 between the census of
1899 and 1929.
1919-1929 773
to 1929 (23,121,589), though of course at a decreasing percentage rate,
depressions affecting the latter only.1 Even in the world crisis and in
the year of minimum registration (1933) the total automobile retail and
service business, including accessories, filling stations, garages, and also
retail sales by wholesalers, figures out at $4,831,800,000.2 Over 1.1 mil-
lion persons were engaged in distribution and servicing, among them
756,000 employees (part-time included), receiving wages and salaries
amounting to 801 millions. Quantitative expansion and qualitative
improvement, falling costs, prices, and rates of profit are obviously the
expected as well as the actual characteristics of this industry's history
during our decade. However, since there is no satisfactory way of
measuring qualitative improvement, and since there was an almost
uninterrupted shift from larger, heavier, and dearer to smaller, lighter,
and cheaper cars — in 1903, for instance, 4.2 per cent of automobiles
produced cost $675 and less, in 1924 nearly 60 per cent3 — even quantita-
tive expansion becomes elusive, while indices of quoted prices, which
should moreover be corrected for variations in the allowances made for
old cars "traded in" and for other forms of rebates, cannot indicate
more than a tendency which, of course, they understate.4 From 1916 on,
profits of individual firms not only fell but also became more nearly equal.
1 Tax-exempt official cars are excluded. Registration of trucks increased through
1980. Possibly a "break in trend" in registration of passenger cars may be said to have
occurred about 1929, after which it may be expected to move over time roughly as popula-
tion in the age groups between 25 and 60.
2 United States Census Bureau, "Census of American Business for 1933." It is of
some interest to note that in the same year state highway expenditure was about 666 million
dollars (United States Bureau of Public Roads), total rural highway expenditure about
1.5 billions.
3 See Epstein, op. cit.t p. 336. Needless to say that does not measure that shift with
any exactness.
4 On this point see Professor Epstein, op. cit., p. 47. He tries to convey an idea by
giving instances of the selling prices of "fairly comparable models," for instance, of a
Packard which sold at $7,000 (without equipment) in 1904 and at $2,585 (equipped) in
1924. The Bureau of Labor Statistics wholesale index of automobile prices falls sharply to
1916, then rises to a peak in 1920, from which it again sharply falls to 1926 — eight points
below the minimum of 1916. Then it rises to 1929, partly at least owing to the reconstruc-
tion of the Ford works. Another index has been constructed by Mr. J. W. Scoville,
Behavior of the Automobile Industry in Depression, (Econometric Society address,
December 1935; published separately), which also displays a trough in 1916. The subse-
quent peak, however, comes in 1918. From it this index falls precipitously to 1923, to
rise to and reach a peak in 1927, after which there is decline, much stronger than in the
case of the Bureau of Labor Statistics index, to 1933. Neither index, of course, overcomes
the fundamental difficulties, which also vitiate cost figures. Attempts have been made to
arrive at a more telling picture by computing prices of cars, or of automotive products in
general, per pound. The Automobile Manufacturers' Association constructed an index
based on the average list prices of the lowest priced five-passenger closed model of each
make, weighted by its relative share in new-car registration. This index overstates the
774 BUSINESS CYCLES
The industry did not simply expand in function of the increase in real
income but helped to bring it about. The former nexus, however,
steadily gained in importance at the expense of the latter, as had been
the case with cotton after the Napoleonic wars and with railroads from
the eighties on. Innovations, increasing in number while individually
decreasing in importance, are typically of the downgrade type. From
1912 on, designs became more stable. Considerable progress in the
standardization of parts and in the rationalization of assembling reduced
costs as did progress in subsidiary industries — tires, nitrocellulose lac-
quers and fast-drying solvents, and so on. Equally important or more
so were the changes in organization and financing that were in part
induced by the struggle for survival within the industry, in which inces-
sant innovating and expanding into the low-price market was a matter
of life and death. Compet ing-down went on at a rapid rate. The rise
in price level after 1916 helped to keep failures and exits at a low and
decreasing figure, and even the setback of 1918, when both production
and wholesale value fell absolutely for the first time, cost few lives.
But after 1921, when production and wholesale value again fell abso-
lutely, exits — not necessarily failures — increased sharply in the midst of
spectacular expansion of the industry as a whole, reaching 21 per cent
in 1924. In 1923 and 1924 no less than 29 firms went out of business,
17 of them war and postwar foundations. Of the 101 plants — makers,
not concerns — whose annual production of passenger cars was 5,000 or
less in 1920 only 11 survived in 1930; of the 23 whose annual production
was from 5,000 to 25,000, also 11; while we still find all of the 10 which
produced over 25,000 in 1920.1 By 1918, 70 per cent of all automobiles
produced in this country and Canada came from the three largest pro-
ducers, by 1921 80 per cent, and by 1935 nearly 90 per cent.
Considering that the car of the masses2 became a reality, while the
industry, which had always been monopolistically competitive, developed
case — it falls by 40 per cent from 1925 to 1932, see chart on p. 6 of Mr. Sloane's message
to the General Motors shareholders of Dec. 31, 1937 — but is, in the particular circum-
stances, perhaps somewhat less objectionable than it would be in general. For the behavior
or profits, see Epstein, op. cit., pp. 243, 256, 264. There it will be seen that the "bonanza
period" lasted well into the Kondratieff recession (1916, according to Professor Epstein),
which conforms to expectation. It is, however, interesting to observe how the leading
firm fared. The ratio of profits to net worth in the Ford concern fell from its 1907 peak
with well-marked cyclical fluctuations heading toward zero until the plant underwent its
complete reconstruction — which amounted to a second, though induced, innovation — in
1927. Number of new companies founded followed those fluctuations with an average
lag of between one and two years.
1 J. W. Scoville, op. cit., p. 24.
2 Tt should be mentioned that the "automobilization" of the farm went further than did
its electrification. At the end of our period about 60 per cent of all farms were equipped
with motor vehicles, and about one-quarter of all trucks were in farm use.
1919-1929 775
a typically oligopolistic situation, we cannot help being painfully aware
once more of the somewhat less than realistic character of the general
conclusions arrived at by the leading theorists of monopolistic competi-
tion.1 In fact, it should be obvious that the behavior of the motorcar
industry during our decade could be described much more convincingly
in terms of perfect competition working under the conditions of a new
industry in the course of being absorbed by or inserted into the economic
system. In the course of this development, ever since about 1916,
methods of financing changed significantly. "Outside capital" began
to play a greater role. We need, however, only mention the direct con-
tact established by General Motors with the open market and its policy
— followed by the other concerns — of financing the consumer. Never-
theless, owned capital accumulated from profits and retailers' and fur-
nishers' credit remained the industry's most important sources of means,
and this accounts for much which strikes the observer as particularly
"sound" about it. Net tangible assets of motor-vehicle manufacturing
plants reached their maximum of about 2.1 billion dollars in 1926 and
then steadily fell, though up to the crisis but slowly. However unreliable
any inference from this may be, it seems clear that, barring the Ford
plant, the great wave of investment belongs to the third and not to the
fourth Juglar.
In order to prove with quantitative precision how much of the proc-
esses of the period and of the behavior of aggregates can be explained by
the motorcar developments alone, it would be necessary to go fully into
what they meant for the steel, copper, and equipment industries and
so on.2 We will, however, confine ourselves to one remark on the
petroleum and another on the rubber industry. Innovations that have
already been mentioned (Chap. VII: flooding, cracking, hydrogenation,
extension of new uses such as fueling of locomotives and ships, by-prod-
ucts) and the discovery and development of new oil fields account for
the fall in gasoline prices (excluding tax) from $0.2411 per gallon in
1919 to $0.1557 in 1929 and — gasoline consumption did not fall until
1 Some exponents of the practical implications of those conclusions — restriction of
output at rising prices and falling profits, uneconomically undersized plants, and so on — •
are, realizing that there is danger that those conclusions might look to us like caricatures,
in the habit of listing the motorcar industry as an exception. But it is only the outstanding
example of a very large class. The tire industry, to be mentioned presently, is another.
Compare what has been said on these points in Chaps. II and X.
2 Even the railroads, which on balance, of course, suffered, did not go without a share.
Transport of vehicles themselves, of steel, gasoline, and lubricating oil contributed a non-
negligible amount to their revenue. This amount becomes even important if we include
road-building material, although the Automobile Manufacturers' Association's estimate
of over 265 millions for 1934 seems somewhat optimistic. Net premia of all types of
automobile insurance amounted to nearly 411 millions in the same year, according to the
Insurance Year Book Service.
776 BUSINESS CYCLES
1932 — $0.1178 in 1931, l which shows that the petroleum industry was
not passively drawn along by the growth of demand. Yet it comes
sufficiently near to this pattern to qualify as an instance. This is par-
ticularly evident at the beginning of the period. In 1920 prices of oil
and gasoline rose considerably (peak of the period), so much so as to
throw them out of line with those of competing fuels and as to restrict
the use of fuel oil by railroads — the Great Northern, for instance, con-
verted 70 locomotives into coal burners. This followed upon the doubling
of automobile production in 1920 as compared with 1918, with which the
gasoline production was then unable to keep pace. An oil boom started
accordingly, which almost coincided with deep depression in other lines.
Issues of oil securities were at a peak early in 1920 and again toward the
end of the year and at the beginning of 1921. It is worth while to men-
tion that the only cities in the country which experienced greater building
activity in November 1920 than in November 1919 were Los Angeles,
Baltimore, and New Orleans, and that the Californian cities all showed
large gains in their clearing figures while these declined in the rest of
the country. At the beginning of 1921 there was a large oil merger
(Barnsdall Corporation). Further developments followed and crude
prices reacted promptly, Midwestern prices, for instance, falling to $1
a barrel in the summer of 1921, as compared with $3 in January. We
need not, however, follow the history of this, in many respects, peculiar
case.2
The rubber industry was, of course, also "drawn along." But its
own innovations were much more in evidence. As we have seen else-
where, beginnings date far back (Goodyear vulcanization to 1839, for
instance) or at any rate to the Kondratieff prosperity (reclaiming, e.g.,
1899, acceleration of the vulcanizing process 1906; but commercial suc-
cess of synthetic rubber came after our period), the use of various pig-
ments in order to increase the durability of rubber compounds (1916)
being the only "inventive" innovation of the twenties. It was again the
"spreading" by means of discovering new and developing old industrial
uses for rubber (flooring, rubber cushions, rubber linings, mountings,
bumpers, and so on) which was a feature of the period under discussion.
In the field of the most important article the great new thing — though
also invented long ago (R. W. Thompson, patented 1845) — was, of
course, the pneumatic tire (1916), which followed upon the success of
1 Figures of the American Petroleum Industries Committee; tank wagon; 50 selected
cities.
2 We may note that in 1927 the industry employed 1.25 million people and that equip-
ment facilities, including pipe lines, were valued at about 11 billions. There were then
830 refineries with a capacity of 8 million barrels crude a day, while the 818,600 wells
produced 2,4 millions.
1919-1929 777
the cord and may be said to have imparted immediately a significant
impulse to long-distance trucking,1 although at as late a date as July 1,
1920, the India Rubber World (p. 633) professed itself unable to believe
that the solid tire, which had greatly improved in reaction to the intru-
sion of the new competitor, would be crowded out. At the same time the
commercial opportunity for low-pressure tires for passenger cars mani-
fested itself in the habit of many motorists to underinflate their tires
for the sake of comfort.2 By 1923, 21 companies, among them prac-
tically all the leaders of the trade, were making such tires, experimentally
or commercially, and several automobile manufacturers had adopted
them as part of the regular equipment of their cars, while others listed
them as optional. A "revolution" in tire making, the more important
because it involved considerable new investment, announced itself.
There was still resistance to overcome. But improvement and stand-
ardization— as to rim requirements — carried the innovation suddenly to
definitive success about 1925, after one of the tire companies had taken
the bold step — in the midst of doubts about practicability and the prob-
able reaction of the public — to bring out balloon tires for all standard
rims and thus to make a bid for immediate replacement of practically
all tires in use. The aspect of the market changed within a few months,
and the "host" followed the innovator promptly. There is no need of
going into the illustrative virtues of the case or the quantitative impor-
tance of it for the fourth Juglar.3 With quick changes in production
functions, the competing-down process asserted itself strongly. We
shall interpret in this sense the symptoms of overinvestment and over-
production,4 observable already in 1923 and again after 1926, and expect
a contribution to the picture of the subsequent crisis from this industry.
1 Registration of trucks increased from 136,000 in 1915 to over 2 millions in 1924, much
more than passenger-car registration. This would hardly have been possible without
the success of the pneumatic tire, first displayed by the Goodyear Wingfoot Express.
2 This instance shows the relation between "wants" and enterprise (in our sense) very
nicely. Consumers by thus clearly indicating their wishes displayed an altogether unusual
amount of initiative, which greatly facilitated the entrepreneurial act. Yet they did not,
off their own bat, "demand" balloons and that practice of underinflating might have gone
on indefinitely without giving rise to what almost amounted to a new industry. This
development was, no doubt, conditioned, but it was as little brought about by consumers'
behavior as was the development from fabrics to cords, supertwists, and rayon cords.
3 Employment in the industry expanded considerably at wage rates which head the
list of basic industries and were above the average of all manufacturing industries. It
has been stated that in 1908 a workman in tire manufacture received on the average 40
cents per hour and could have bought for $35 a tire which was good for about 2,000 miles,
while in 1936 he received 88 cents and was able to buy a tire which cost $8 and gave 20,000
miles of service, so that an hour's labor gave him about 95 times as much tire service in
1936 as in 1908.
4 E. G. Nourse and Associates, America's Capacity to Produce, 1934, pp. 236 and 237,
estimate average utilization of capacity for 1925-1929 at 85.3. This is high under the
778 BUSINESS CYCLES
c. The heavy chemical industry had, as we have seen, developed well
before the war, but enterprise in the organic branch was entirely condi-
tioned by the seizure of German patents and later on by protection.1
Prices of chemicals, which according to the B. L. S. index (1926 = 100)
were at 89.4 in 1913 and which had, owing to the practical cessation of
German imports, soared to 197 by 1916, testified to the vigor of entre-
preneurial response to those new conditions by falling to 97.2 in 1922.2
Both the coal-tar group in all its stages, particularly in the production
of dyes, and the aliphatic group scored a series of successes that extended
over the whole of our period and throughout the subsequent depression
and amounted to the creation of new industries. Investment, employ-
ment, wage bill, profits, and dollar volume of sales — about 2J4 billions
towards the end of our decade — increased to a peak in 1929 for the chemi-
cal industry as a whole. Sales in the non-coal-tar group continued to
increase without break afterward. Medicinals, solvents, perfumes,
antifreezes, carbon tetrachloride, acetic anhydride, camphor, resins,
nitrates (synthetic iodine and synthetic rubber came early in the thirties)
may serve as examples. Analysis of the individual cases would show
little more than so many instances of the way in which innovation works.
Three points only call for additional comment. First, for the same
reasons as elsewhere concentration of control and research and coordi-
nation of specialized large-scale plants were in evidence in this industry.
The Dupont concern, like the J. G. Farben, expanded far beyond the
chemical field.3 The other giant, the Allied Chemical and Dye Corpo-
ration, was the result of a merger in 1920 of five big concerns — three of
which were in the *' heavy" field — which to a large extent complemented
each other. Second, new branches of industry emerged around what
circumstances and reflects the increase in production from about 2 million tires in 1918 to
nearly 8 million in 1929. But it is perfectly compatible with considerable underutilization
and overproduction (the latter in the sense that "old" firms often failed to cover their
costs) in spots.
1 Developments illustrate, however, the rationale of the distinction between condition-
ing innovation and innovation itself. Whoever feels inclined to doubt it is invited to try
to deduce these developments uniquely and ceteris paribus from those two events.
2 Then they rose to hover around 100 (there was an upward movement in 1925) through
1928, when they embarked in consequence of quantitative expansion at falling costs on another
downward course. It is important to note, first, that this again affords an illustration of
popular statements about rigidity and, second, how much more obvious is the influence of
individual prices, falling under the impulse of innovation, on the price level than any
influence of autonomous monetary factors acting through the latter on individual prices.
There should be no uncertainty about the diagnosis and the significance of such facts
that say indigo fell from $2 a pound in 1916 to 14 cents in 1982, or barbital and veronal to
about one-seventh and aspirin to about one-eighth of its prewar price.
3 The interest taken in General Motors is one example of this method of acquiring and
buttressing markets.
1919-1929 779
may be called the production of chemical fundamentals. A host of small
and medium -sized firms took up the production of a truly unsurveyable
variety of drugs, cosmetic articles, and so on. The results, as distin-
guished from the formal properties, of monopolistically competitive situ-
ations are much more in evidence in this group and within its army of
retailers and advertisers than among the few big producers of the basic
stuffs. For us it is important to note the quantitative importance of this
trade and to account for its spectacular expansion: the unrivaled oppor-
tunity which it exploited was one of the consequences of the increase in
the real income of the masses which left even the lowest income groups
with a surplus that was not a priori allocated to specific purposes but
ready to go wherever advertisements beckoned. The phenomenon thus
fits well into our ideas about downgrade developments. Third, the
chemical industry displays the (secondary) competing-down process
within the innovating line much less than, say, the automobile industry;
but it displays the (primary) competition, i.e., the competition with other
commodities or older methods of producing the same commodities much
more than almost any branch of economic activity. In some cases its
innovations act through other spheres of production, agriculture for
instance. In others they act directly and then with a promptness to
the consequences of which the social fabric of capitalism may well prove
unequal some day. Chemistry provides not only acceptable and cheap
substitutes for things that are the basis of much employment and invest-
ment, but quite often exactly the very same things — frequently in a better,
especially more uniform and more reliable quality, as for instance in the
case of varnishes and dyes — which had been produced by nonchemical
methods before. It does so almost always at a cost which eventually,
though not as a rule immediately, falls far below the level attainable by
the latter. In such cases large sectors of the economic organism may
have to go out of operation at very short notice. If the consequences
have so far not made themselves more strongly felt in our three countries,
this is because they mostly impinged on others, on Chile for instance or
India — or in the case of madder, on the countries from Southern France
to Asia Minor — or Sicily (citric acid, 1927) or pro futuro on the rubber-
producing countries. The United States, England, and Germany were,
during the period under survey, not much affected in this respect, and
whatever effect there was was rather favorable. But more serious dis-
locations may arise from such developments some of which are obviously
imminent. The term revolution acquires in this connection a particu-
larly ominous meaning. Depressive influences may emanate from this
line of advance by comparison with which anything that can be effected
by action on monetary aggregates, central bank action included, is of
negligible importance.
780 BUSINESS CYCLES
The rayon industry, of course, owed much to the tariff, and its great
concerns owed much to their control of patents. But in all other respects
the case is strikingly analogous to that of the automobile industry. We
have a sharp competitive struggle at the beginning of the period, partly
due to the numerous short-lived foundations after the war, from which,
as has been stated elsewhere, emerged three concerns which accounted
for about 90 per cent of the production of the country. In this oligo-
polistic setup the great expansion of consumption took place1 which was
but little affected by the world crisis. Wholesale prices (150%denier, A
grade, New York) fell from the 1918 peak to about the prewar level
(1914, $1.96 per pound) in 1925, and were at $1.25 by 1929.2 Profits per
pound of product steadily declined, although in the case of the American
Viscose, which remained the leading producer throughout, they were
still 58 cents in 1928.3 Other textiles, though not without some propelling
influences — higher cotton consumption per spindle, production of cord,
artificial leather, broadcloth shirting, fancy woolens and so on — behaved
like the old industries they were. Quantitative expansion and qualita-
tive improvement were considerable, and there was much rationalization
in details. This does not alter the fundamental traits of the picture,
which are reflected in the behavior of prices.4 Continuing locational
shifts caused as much sectional depression as sectional prosperity. The
Department of Labor's combined index of employment kept steady
throughout the decade, but nevertheless marks strongly the upswing
that set in during the second half of 1925.
d. As far as changes in production functions go, the iron and steel
industry should really be dealt with — as should metallurgy in general —
in connection with chemistry and electricity. It suffices to mention
the career of light alloys, the first stage of which was run during our
decade, especially from 1925. There were also technological and organi-
zational changes of other types, such as continuous rolling or the crowding
out of the merchant furnace, and of course many improvements of the
1 Deliveries — shipments of American firms plus imports minus exports — as computed
by the Textile Economics Bureau and published in the Rayon Organon and elsewhere,
increased nearly threefold from 1923 to 1927 and again by about 50 per cent to 1931. They
do not include acetate rayon, which gained ground steadily and was in 1925 little more
than 3 per cent of the total production, but 7 per cent by 1929 (nearly 22 per cent in 1935).
2 What has so far been the minimum, 50 cents, was reached in April 1933. Here again
the contribution to the price level of the crisis period is more significant than the contribu-
tion of that level.
8 The oligopolistic pattern developed into a monopoloid one through the merger in
1925 (American Rayon Production Corporation).
4 The B. L. S. group index for cotton goods (1926 = 100) was 56 in 1914, 147.5 in 1919,
and 190.7 in 1920. Then it fell abruptly, but was still 98.8 in 1929. Only in 1932 it fell
below the 1914 figure. The group index for woolens and worsteds behaved similarly.
1919-1929 781
rationalization kind in individual lines or concerns. Increasing use of
scrap in the steelmaking as well as in the copper, aluminum, and other
industries deserves particular emphasis.1 Speaking broadly, however,
the steel industry suffered in depression — especially in 1920 and 1921 —
and prospered in booms — the peaks in pig-iron output occur in 1923,
1925, and 1929 — in consequence of the general business situations, rather
than in consequence of its own enterprise. The behavior of its prices2
accords with this impression. Steel consumption increased strongly,
however, in spite of all the steel-saving rationalizations, which were
more than offset by the conquest of new uses — steel increasingly became
a consumers' good — and general expansion. Per head of population it
was, at the end of our period, seven times as great as it had been in 1900.
Nothing fundamentally new happened in the aluminum and copper
industries. We have, however, already observed in an earlier place that
the former displayed great initiative in discovering and conquering new
ground. Its quantitative expansion — domestic primary production
more than doubled between 1915 and 1929, while domestic consumption
increased more than threefold, secondary production accounting for the
greater part of the difference — was one of the major industrial features of
the period, another good instance of a downgrade development. Price
behavior was in accordance with this and characteristically different
from that of finished steel. The absence, comparatively speaking, of
fluctuations around the fundamental contour affords an interesting
example of what " control " by one firm really means under conditions
of rapid growth and of commodity competition. The domestic price
of new aluminum ingot 99 per cent pure reached its war maximum by
1916 and its prewar level or, corrected by the B.L.S. index of wholesale
prices, a figure nearly 30 per cent below it, by 1922. It then slowly rose
to 1925, afterward fell somewhat, and was maintained at 23.3 cents
1 This material-saving practice, a typical downgrade development and responsible
for the increasing divergence between pig-iron and steel outputs, of course exerts a depress-
ing influence on the production of a number of important raw materials also outside the
metal field, and constitutes in each case a distinct innovation as well as a distinct industrial
problem. The prices of scrap are more sensitive to the course of cyclical phases than any
other commodity prices and, as has been pointed out by the Berlin Institute (e.g., in 1926),
the relation between scrap- and pig-iron prices is a good index, even forecaster, of business
situations.
2 Steel prices never came down to the 1914 average and kept substantially above it
even during the crisis. Their behavior is interesting because of the differences in the price
policy of different concerns, which make it difficult to speak of one price at a time. The
lead of the United States Steel Corporation was not consistently followed by the inde-
pendents. But neither did the latter do what one might expect from younger and more
vital concerns, viz., take an initiative in reducing prices. On several occasions they were
more eager to advance and more reluctant to reduce them than the United States Steel
Corporation was.
782 BUSINESS CYCLES
through 1934. 1 Thus it failed to fall in 1930, when it would have done
so under competitive conditions. But the profits made are not in them-
selves sufficient to prove that in the long run prices were, given the pro-
tection, substantially above — they even may have been below — that
level at which they would have moved had competitive conditions pre-
vailed from the outset, provided we include in the latter the degree of
productive efficiency compatible with the competitive scale of individual
firms. Nor does it follow that, if all or most industries had been organized
in the same manner, they would have still found it to their interest to
adopt the same policy of price stability.2
The war for obvious reasons brought a large expansion in the con-
sumption of copper, to which the new mines and mining methods (see
Chap. VII) were, however, fully equal, so much so that already in 1917
efforts were made to fix prices. They were followed by others throughout
the twenties, which in fact succeeded in keeping price fairly stable at
about prewar level (12 to 15 cents) from 1921 to 1928. At the same time
costs were being incessantly reduced by further development of large-
scale mining methods and by new processes of smelting and refining as
well as by the discovery of better deposits. There were also important
horizontal combinations — among mining companies, smelters and refiners,
and copper- and brass-producing concerns — and vertical ones. Thus
an untenable situation developed — indicated by the fact that production
kept persistently above consumption, stocks being well above prewar
level throughout — under what for a decade seemed a prosperous surface :
one of the weak spots that were to contribute their share to the processes
of the world crisis. Secondary copper — in 1929, production of secondary
copper from scrap was 40 per cent of smelter's production of new copper
from domestic ore — and the output of low-cost mines in Canada, Katanga,
and Northern Rhodesia did the rest. The formation of an international
export cartel under the Webb-Pomerene Act (Copper Exporters, Inc.,
October 1926, which "controlled" 90 per cent of the world's production)
only deferred the catastrophe.
We have here an extremely interesting case of an otherwise perfectly
unfettered process of capitalist innovation, which was interfered with
only by capitalist interests themselves and by these again only through
an attempt to put out of operation a single element of the mechanism,
1 Data are from the Engineering and Mining Journal. The American Metal Market
gives for the same year an average of 21.6.
2 In that case reduction by agreement of all those prices would have been practicable
and might have suggested itself because, with respect to all of them, demand would not
have been as inelastic as it was with respect to each of them taken singly. In the case of
public price fixing, on the other hand, that stability might have become an end in itself,
as it did in most cases of price regulation by public authority.
1919-1929 783
viz., the effect of "progress" on price. It is worth while to consider
what the course of events might have been without such an attempt.
Prices would certainly have fallen and it may well be that especially in
the short run this would not have increased consumption appreciably.
But this is not the point. Mines and refineries which actually were
kept alive would have been eliminated in any case, though of course
more of them if demand was really inelastic in the relevant range than
if it was not. And this would have first eliminated waste — for it is
social waste to work a mine or refinery which can be worked only at an
"artificial price" — and second helped to tone down the prosperities of
the twenties, to spread the work of readjustment and pro tanto to mitigate
the subsequent crisis. If all this be impossible or more than the system
can stand, public regulation or ownership is, under such conditions —
the reader should carefully analyze in what they differ from those of the
aluminum case — in fact the only alternative to violent breakdowns,
though not necessarily a remedy for this type of economic waste.
e. While it is believed that the above exposition brings out the salient
features of the industrial processes of the decade under discussion, estab-
lishes what we have set out to establish, and suffices for our purposes —
particularly for the purpose of interpreting the business situations of
the period and the behavior of aggregative time series — it must once
more be emphasized how very incomplete it is. It even leaves out some
major elements — natural-gas-pipe lines (1927) have, for instance, not
even been mentioned — and practically all those minor ones1 the sum
total of which is particularly important in the downgrade of a Kondratieff .
Knowledge of the full extent of the revolution which that period wit-
nessed both in the methods of production and commerce and in the
structure of the budgets of households, and an adequate analysis, in the
light of it, both of the period itself and of its aftermath, would presuppose
very many case studies beyond those we at present have.2 Nevertheless,
the main features stand out unmistakably and can be further illustrated
by a few facts from the Abstract of the Census of 1930. 3 This census
1 A major movement, which however resolves itself into an almost infinite number of
small ones, is what may be called Taylorization. Its spread during our period is a typical
consequence of the struggle for survival amidst the readjustments of downgrades. The
pressure of this country's wage level adding momentum, this type of rationalization of
every job was in many cases more effective in reducing costs per unit of product than
fundamental innovations could have been — and in all cases highly significant from our
standpoint. It is a special case of a class of which the efforts to utilize scrap and waste
are another.
2 The most important achievement in that field is Mr. H. Jerome's Mechanization in
Industry, herewith recommended once more to the attention of the reader as a storehouse
of examples of typical downgrade improvements by which to test our interpretation of the
period.
» Pp. 744-759.
784 BUSINESS CYCLES
lists 103 industries each of which had in 1929 a Value Added of over
50 million dollars and also was independently listed in 1919. On the
average (unweighted) Value Added increased between these two years
by 29 and the ratio between Value Added and Pay Roll by 16 per cent.1
First, however, we are interested in those industries which, while pro-
ducing a Value Added of over 50 million dollars in 1929, do not inde-
pendently occur in 1919, since this in itself proves a very rapid rate of
growth. There are 16 of them: beverages,2 food preparations, millinery,
motion pictures (excluding projection in theaters), paper, pulp, rayon,
refrigerators, rubber tires and inner tubes, other rubber goods (excluding
boots), and typewriters being the most significant ones. Besides repeat-
ing cases which we know already, this list adds a few new elements to
our picture. Of particular significance is the suggestion, which under-
lines one feature of that phase of the long cycle, of industries which
expanded simply in response to the increase of consumers' real purchasing
power and without any particularly strong impulse of their own.
This suggestion grows stronger still if, second, we glance at those
among the 103 industries that display an increase in Value Added of,
say, more than 100 per cent. Besides aircraft and parts, which heads
the list (510 per cent), electrical machinery and supplies, aluminum
manufacture, motor vehicle bodies and parts, which we would expect
to find, we also meet perfumes and cosmetics, signs and advertising
novelties, concrete products, flavoring extracts and sirups, photo-en-
graving not done in printing establishments, house-furnishing goods,
ice cream, printing and publishing (newspaper and periodical). Patent
medicines, soap, cigars and cigarettes, cereal preparations, bakery prod-
ucts, while not reaching the 100 per cent line, yet increased their Value
Added by much more than the average figure so as to reinforce the evi-
dence. Third, we will note some of those industries the Value Added
of which decreased by more than 10 per cent: shipbuilding, locomotives
(not made in railroad repair shops), railway cars, pianos, phonographs,
leather, beet sugar (Value Added of the cane-sugar industry remained
at the 1919 figure), cotton, woolen and worsted goods, flour.
rThe limitations of Value Added as a measure of the development of an industry,
and of the ratio between Value Added and Pay Roll as a measure of innovation or induced
and completing rationalization are too obvious to be explicitly stated. It is easy to see,
however, in what sense both are relevant to our subject.
2 The industry of nonalcoholic beverages was, of course, conditioned by prohibition.
There were, however, several minor innovations, some of them of a purely commercial
type. But prohibition also conditioned the creation of organizations for the illicit trade
in alcoholic beverages, which constituted innovations and a new industry in our sense.
The somewhat unconventional character of this industry only serves to illustrate some
aspects of our concept of entrepreneurship. If we exclude it, this is only because the writer
had to come to the conclusion that data were too unreliable for use.
1919-1929 785
We cannot expect, nor do we find, significant correlation between per
cent increase in Value Added and per cent increase in the ratio of Value
Added to Pay Roll. Also, some new or relatively new industries, such
as motor bodies or aluminum manufactures, understandably show very
little signs of the effects of labor-saving devices on that ratio, although,
of course, others, such as motorcars or aircraft, rank high, and some of
the old and conspicuously noninnovating ones, such as house-furnishing
goods (1 per cent decrease), rank low in this respect. It is, however,
instructive to observe — and tells a great deal about the general character
of the industrial processes of that time — how much labor-saving rationali-
zation went on outside of the great lines of innovation. Thus the ratio
between Value Added and Pay Roll increased by 120 per cent with
cigars and cigarettes, 85 per cent with soap and with coke (excluding
gashouse coke,) 71 per cent with cereal preparations, 61 per cent with
manufactured gas, 52 per cent with cutlery and edge tools, 52 per cent
with ice cream. Even for tin cans that figure is still 33; for patent
medicines, druggists' preparations, coffee roasting and grinding 32;
cane-sugar refining 28; meat packing (wholesale) 26; butter 26; cement
and concrete products 22; perfumes and women's clothing 17. Only in a
minority of cases — printing would be one — was this due to substantive
novelties that we have simply been unable to mention. In the main it
was the result of that systematic effort to fight, under the pressure of
a price level that tended to fall, each cost item by exploring every detail
of the productive and the commercial process and by applying and devel-
oping techniques the fundamentals of which were fully established before
the war, but which in many cases involved not only technological improve-
ment in existing plant but also the erection of new, highly mechanized
plant.1
/. As pointed out elsewhere, it would for this country be possible to
carry our count of Juglars and Kitchins through the war, which never
succeeded in blotting them out completely. But in order to avoid a
statement which cannot be usefully made unless it be more fully estab-
lished than is possible here, we will now not go beyond saying that the
end of 1916 and the beginning of 1917 might, but for the war, have wit-
nessed what according to our terminology would have been the beginning
of the prosperity phase of the third Juglar of this Kondratieff, and that
in this case the crisis of 1921 would have occurred exactly when our
schema would have led us to expect it, i.e., when that Juglar turned from
1See E. F. Baker, Unemployment and Technical Progress in Commercial Printing,
American Economic Review, September 1930, and Technological Change and Organized
Labor in Commercial Printing, American Economic Review, December 1932. The great
increase in our ratio under the heading Cigars and Cigarettes was in part due to the new
machine introduced into cigar making.
786 BUSINESS CYCLES
recession into depression. Even if we wished to press this — which we
do not — it would leave us all the freedom in the world to take into account
the effects of war demand and war finance, of the shock imparted to the
war structure by the armistice — i.e., the four months of dullness or
wavering rather than collapse which followed upon it — of foreign and
domestic postwar demand producing the boom of 1919 and, finally, of
all that partial liquidation of both war and postwar situations contributed
to the slump of 1920-1921. * To what has on various occasions already
been said about the latter it is necessary to add but two comments.
First, however clear it is that that slump was primarily a process of
liquidation of war effects and a reaction to the boom of 1919 — which in
turn had little if anything to do with innovation — yet the presence, and
businessmen's awareness of the presence, of a new industrial situation —
a situation which was new in the sense that the consequences of prewar
innovations had profoundly altered the cost structure — had much to do
with the severe restriction in output of manufactures which first began in
January and again, after a rebound, in March 1920, in the face of the
facts that retail sales had throughout 1919 increased at a greater rate
than had production,2 that the export trade as yet showed no signs of
slackening, that prices continued to rise. Banks, moreover, were, by
the influx of gold (gold imports in 1921 amounted to nearly 750 millions)
and by the reduction of government debt by about 1.2 billion dollars
between June 30, 1919 and June 30, 1920, enabled to increase their loans
by about 1.5 billion dollars during the same time.3 All this puts some
of the most popular theories out of court in this case. We have once
more an instance of "business deflating itself" without any serious
outside pressure, and we see again that this could have been prevented
only by continuing government expenditure at the war level or a level
still higher. The question why business deflated itself cannot be given,
even in this case, without reference to our mechanism.
Second, the reaction was sharp and unimpeded and, because it was
sharp and unimpeded, relatively short. Prices and wages were allowed
to drop drastically, liquidation of commodity stocks and debts proceeded
1See W. M. Persons, The Crisis of 1920 in the United States, American Economic
Association, 84th Annual Meeting (December 1921). The picture which that paper draws
of the situation at the time is perfectly sufficient for our purpose, so that this reference may
relieve our text of the necessity of going into details.
2 This has been emphasized by Professor Slichter, The Period 1919-1936 in the United
States, Review of Economic Statistics, February 1937. Although considerations of space
forbid the presentation of the year-to-year analysis of business situations from 1921 to 1929,
carried out for the purposes of this book by Professor E. M. Hoover, comparison between
Professor Slichter's business history of the period and the remarks that follow in our text is,
nevertheless, invited.
3 Federal reserve rediscounts reached a peak of 2,827 million dollars on Nov. 5, 1920.
1919-1929 787
rapidly, elimination of firms — over 8 per cent of the manufacturing firms
which were in business in 1919 had disappeared by 1921 — was prompt,
money rates fell, credit was readily available, and the situation began to
stabilize itself in April 1921, the textile and clothing industries, which
had expanded first in 1919 and fallen first in 1920, being among the first
to revive.1 The resulting price relations differed greatly from those of
1913 and struck many observers as entirely abnormal. But the change
was largely, though not wholly, one of adaptation to new conditions.
Though improvement slackened in October and many signs of con-
tinuing liquidation — e.g.y further reduction in wages — outlasted the sum-
mer of 1922, "deep depression" was over by December 1921. In April
1922 the automobile and tire industries experienced shortage of labor,
while stock issues had already revived in January: it was then that the
boom in public utility stocks began. Prices of steel, tires, glass, and
oil rose in the fall, while those of gasoline, automobiles — the price of
tractors had been reduced before by action of the Ford concern — cement,
and foods fell. In December 1922 the oil industry was breaking all
previous records in output. The fact that such should have been the
situation only one year after what nobody doubts had been a major
crisis and in the presence of many depressive symptoms — federal reserve
rediscounts, however, which fell to a trough of 380 million dollars can-
not, owing to the gold influx, be counted among them — is full of potential
lessons which are as obvious as they are useless. The case also shows
better than any theory could how the system pulls out of troughs under
its own steam and how it succeeds in doing so while price level is still
falling.
Our diagnosis then, which it would take very strong preconceptions
to doubt, is simple: abnormally short depression phase of the Juglar,
lasting from the fall of 1920 to July of 1922, owing to abnormally effec-
tive liquidation. Alternatively, we could express the same facts by
saying that the depression phase lasted to the end of 1922 but that its
work had so effectively been done by May 1921 that, the ground being
cleared, the prosperity phase of the third Kitchin, which, as we know, still
belongs to the depression phase of its Juglar, had unusual opportunities
of asserting itself. In any case this Kitchin stands out unmistakably,
and there is no reason why we should not so call an undoubtedly short
1 Of course, the above statement excludes those commodities which, like oil, electricity,
gas, meat, motors, had either not fallen at all or fallen only a little. The activity in the
petroleum industry has been emphasized before. Several automobile factories which had
been shut down reopened in April. The tin can and cigar industries (new plant of the
Continental Can, activity of the Bayuk Cigar Company) were also active. So were the
shoe and leather and the drug industries. One cigarette company reported that in the first
months of 1921 it had earned more than double the whole year's dividend requirements.
788 BUSINESS CYCLES
cycle which is universally recognized — even if differently dated by stu-
dents who count from trough to trough — and which completely answers
expectation from our schema both as to formal characteristics — length
and location included — and as to industrial meaning. On the other two
Kitchins which within that Juglar ought to have preceded this one, we
will not insist, although it would not be difficult to establish them
statistically.
What followed, from either the middle or the end of 1922 on, either
is or very much looks like a normal Juglar recovery, which lasted to the
autumn of 1925. Our schema would lead us to expect that it contained
a setback owing to the Kitchin depression which it tells us should have
occurred. It did occur. After it had run its course, recovery resumed
and from August 1925 on both Kitchin and Juglar were shading off into
the prosperity phases of what then would be the fourth Juglar and its
first Kitchin.1 A few additional facts may be useful. In the first
quarter of 1923 the upswing reached its peak. Unemployment was at
low ebb — some people spoke of its being "absent" — in February. Most
prices, especially those of metals and building materials, tended upward;
a record year for construction was correctly foreseen. Steel — 19 new
steel furnaces were built in 1923 — coal, and cotton textiles expanded.
Four new power stations were announced for construction. Railroads
"came into their own again" and gave orders. Everything except agri-
culture and ship-building boomed. Wage rates rose strongly. April
saw record figures but also a break in the stock market. Signs of slack-
ening activity began to show by August, attributed as usual to external
factors, and by December expectations were not very optimistic. They
were borne out by the state of things during the spring and summer of
1924. More important than the uncertainties incident to the presidential
campaign were the — understandable — reactions in the automobile and
oil industries. Steel consumption, railroad traffic, employment, and
prices fell.2 Nothing very serious happened, however; failures of com-
mercial firms were but insignificantly above the 1923 figure; residential
building kept up well, and so did power production, the radio industry,
and other lines. The first two months of 1925 were disappointing — with
employment in most industries below what it had been a year before —
and a collapse of the stock market followed in March. During the second
quarter business was described as steady but "spotty." New financing
and other indications of imminent prosperity asserted themselves under
this surface, however, and, with the help of improvement in the agrarian
1 Professor Hoover wrote in the report already mentioned: "By August the upsurge was
unmistakable and the word prosperity began to creep modestly back into the vocabulary
of the [Financial and Commercial} Chronicle"
2 There was an abortive rally in November.
1919-1929 789
sphere and the — largely speculative — land boom ("real estate subdi-
vided"), the fall wore on amidst record investment, bank-clearing
(October) and construction figures, rising money rates and steel prices,
stock market excesses (October), failures at record low.1
The explanation of all this — the "ignition" — will be found on referring
to the above survey of the fundamental industrial processes of the period.
They clearly change during those months owing to the influence of several
new impulses — while others, such as residential bulding, lost force — and,
by conforming exactly to what we mean by a Juglar prosperity, justify
our dating. So much is provable and indeed obvious. But we will
for a moment trust our schema to the point of absurdity and try to
"predict" the subsequent business situations on the assumptions, first,
that the fourth Juglar started with the fourth quarter of 1925 — although
we know that no such exactness is possible in historical analysis — second,
that its duration was to be exactly 9^ years — roughly the average
duration of prewar Juglar s — and the duration of its Kitchins exactly
38 months; third, that all the Juglar and Kitchin phases were of exactly
equal length. This absurd experiment yields the following results: the
Juglar prosperity lasting into February 1928 should be interrupted by a
Kitchin depression from May 1927 to the middle of February 1928;
and the recovery of this Kitchin — to the end of November 1928 — and the
prosperity — to the middle of September 1929 — and the recession of the
second Kitchin should then run their course within the Juglar recession
ending with June 1930. At that date both the Juglar and the Kitchin
should enter upon their depression phases on a Kondratieff that had
already entered upon its own,2 so that the configuration of 1873 would be
repeated. The reader will realize that no value attaches to, and no
significance is claimed for, the exact dates. But he will also realize that
the absurdity stops at the assumptions which are responsible for the
exact dates. Stripped of this unwarranted exactitude and confined to
essentials, the "predictions" of the schema are not absurd but on the
contrary tell several important truths — and not a single untruth. It
1 Mr. Thorp's Annals closing with 1925, we may compare his description with ours.
Bearing in mind differences in terminology, we find almost complete agreement. He notes
what above was described as dullness in 1918 ("recession") and the boom of 1919. 1920 is
by him described by prosperity-recession-depression, 1921 by depression. 1922 is a year
first of revival then of prosperity, which is obviously meant to convey exactly what was
meant above. In 1923 there was prosperity-recession and in 1924 mild depression-revival.
1925 Mr. Thorp describes, as we should expect he would, as a year of prosperity without
qualification. The present writer is not sure whether employment was really "full" or
that the rise in bond prices accords with the pattern of prosperity. But there is no material
difference in diagnosis.
2 It will be remembered that according to our schema Kondratieff depressions begin
with Juglar prosperities. The depression phase of the current Kondratieff would, hence,
date from the fall of 1925.
790 BUSINESS CYCLES
should be added that the comparative severity of the setback in 1927,
which was to occur and did in fact occur within the prosperity phase
of the Juglar, does not in itself run counter to expectation : the depression
of a Kitchin, located as that one was, after the end of the Kondratieff
recession, should be well marked.1 On the other hand, the boom of
1928-1929 was more violent than our schema leads us to expect, which
in explanation has but a Kitchin recovery and a Kitchin prosperity to
offer. This may, however, be accounted for by certain autonomous
monetary factors and the influence of the speculative mania, of which the
first do not form part of our model and the second — also present in 1872-
1873 — is always an irregular factor.
The stock market suffered collapse in February 1926. But this is
merely a normal incident of a Juglar prosperity outgrowing its initial
stage. A no less regular phenomenon was, on a Kondratieff depression,
the tendency of prices to sag. If business conditions began to display
signs of "relaxation" already by April, when automobile concerns did not
do so well as they had a year before and the cotton, silk, sugar, and other
industries headed toward curtailment, and if in May there was also a
decline in steel production, this is sufficiently accounted for by previous
speculative excesses, in particular by the passing of the real estate boom.
The stock exchange recovered by June, and almost everything was again
at prosperity levels by August, motorcars and textiles included, furniture
enjoying record profits. Oil developments in California, in North Texas,
and on the Gulf Coast did not entail any large increase in stocks. By
October, however, the 1927 setback came definitely in sight. The stock
market discounted it, bank debits were running from 5 to 9 per cent
below the figures of the preceding year, demand for steel dropped until
operations were at 65 per cent capacity. Failures increased. Car load-
ings also were at the end of what nevertheless was a record year at a
lower figure than they had been in 1925. Some anxiety was felt about
installment sales.2 The agrarian situation had also become more
unfavorable.
1 Professor Mitchell dates one of his cycles from 1927. Since the writer naturally
wishes to differ as little as possible from so outstanding an authority, he begs to emphasize
that no difference in diagnosis of situations is implied in such dating, because it is simply
the consequence of Professor Mitchell's principles to count from trough to trough and to
recognize but a single type of cycles. The particular pattern under discussion seems well
qualified for suggesting that some important elements of reality are being missed if we put
troughs such as occurred in 1924 and 1927 on the same level with troughs such as occurred,
say, in 1908 and 1921, and these again on the same level with the troughs of 1875 and 1932,
and that the distinction of cycles of different type seems the natural way of recognizing
these very real differences.
2 Consumption was at a high level. But no less than 10 per cent of the cars sold
during that year could not be paid for and had to be repossessed by the dealers — a good
illustration of the theory of underspending or oversaving.
1919-1929 791
Until May 1927, however, general business kept on a high level in
spite of all that and even improved, several new things — the Chevrolet
and Frigidaire successes, motion pictures, the North Carolina power
plant, a number of smaller events — supplying impulse. In April busi-
ness was prosperous. But then a definite decline set in — which we may
identify as a Kitchin depression — intensified by widespread recognition
of unsound practice in many fields, with retail and wholesale trade at a
lower level, many failures, and cautious reserve in large-scale business.
Building, the condition of which was complicated by the liquidation of the
Florida boom, was a particularly weak spot. The fall in automobile
production was, of course, due to the reconstruction of the Ford plant.1
The Mississippi flood, while it also explains some things about the behavior
of physical indices, has in other respects to be listed as an impetus. There
was no slump. Good business in the cotton, rayon, and shoe industries
and a continuing stream of new things — Dicsel-engined locomotives,
gas-pipe lines, the refrigerator merger, development of the Kraft paper
industry in the South, radios — were features throughout. By Decem-
ber improvement was almost general, although employment in building
was still 12 per cent below that of a year before.
Railroad earnings, steel production, and gasoline markets improved
in January 1928, and the "bankers' loans boom" in the stock market
was getting under sail with automobile, copper, and rubber stocks leading.
But the general business situation2 behaved until March (even in March
there was considerable unemployment) in a manner which is in our
terminology not badly rendered by the phrase "conditions of Kitchin
depression relieved by a Juglar prosperity." In April, however, steel
was at record rates. So was tire production and by June everything —
building, the automobile (contraseasonally) and the oil industry included.
1 But the writer fails to understand how some observers could have attributed the 1927
depression wholly to that fact.
2 The contemporaneous comments of the public press, and business opinion as revealed
by public utterances of executives and so on, quite correctly realized and sized up some of
the fundamental elements of that situation. The writer's impression is that there was not
only more insight but also more foresight in considerable sectors of the business community
than is in general realized, and also more than is compatible with some theories which
overstress the element of error in the explanation of depressions. But one also finds the
tendency, especially in press comments the authors of which aim at semiscientific explana-
tions, to overstress external and banking factors. The preceding note affords an example
of the former type. The latter type is instanced by the emphasis on Federal Reserve
Board policy in explanation of the state of things in the first two months of 1925 and again
the first two months of 1928. In both cases the industrial process itself supplies much
more plausible explanations. But it is easy to understand how tempting that explanation
was, especially for writers whose contact with business was primarily through the money
market and who were strongly influenced by the scientific and pseudoscientific "theories"
of the epoch.
792 BUSINESS CYCLES
Symptoms of "high" prosperity then went on intensifying themselves
until October, when mail-order sales broke all records. Construction of
new plant — for old as well as for some new purposes — new financing —
taking advantage of the stock-market boom — dividends, money rates,
and so on were all in keeping with the rest of the picture, into the details
of which we are unable to go. There were two apparently discordant
elements, however. First, unemployment increased. Second, com-
modity markets though buoyant were not really sellers' markets: the
almost desperate efforts made by the sales organizations of big and small
concerns and the fact that such increase in prices as occurred was insig-
nificant, while many important prices had to be reduced, indicate a
certain strain in the system.
Now if the reader will remember the writer's various attempts to
convey his idea of a Juglar recession, he will appreciate the warrant for
expressing that state of things by saying that it was exactly what we
should expect from a Juglar recession on a Kondratieff depression coupled
with the two positive phases — strictly, according to our experimental
schema, the recovery phase only — of a unit of the short cycle: good and
expanding business accompanied by increasing unemployment and by
that strain which is the consequence of the "avalanche of goods" smash-
ing its way through the resisting framework of the existing industrial
structure — this is precisely the picture which would result from that
particular juncture. As stated before, however, there is no doubt that
the developments between April 1928 and August 1929 added to the
situation many of the untenable elements1 which subsequently served
to intensify the crisis.
October 1928 brought the first symptom of slackening activity, which
was, however, to disappear temporarily by January 1929: with the
exception of Ford, all automobile producers then decreased their pur-
chases of steel.2 In November total building fell off more than seasonally.
But barring building and production of building materials, which con-
tinued to decline, most lines of industrial and commercial activity sur-
passed 1928 output figures during the first six months of 1929 at (from
September 1928) falling prices but high profits. Also plant construction
and financing seemed to have taken out a new lease on life. Quite a
list of new things (at least of the "induced" type) were being inaugurated
in June, when pig-iron production reached a maximum. Moreover, the
1 This question will be touched upon in the next section. But it is clear that whatever
it was that may have to be explained by either the mania or monetary factors or both
together superimposed itself on, and made "excesses" of, "normal" phenomena of the
same nature.
2 The effect on the steel industry was, however, mitigated by a simultaneous increase
of demand from the oil industry.
1919-1929 793
aircraft, radio, refrigerator industries prospered. So did automobiles,
tires, machine tools and other implements, hardware, cotton, silk, rayon,
and cigarette production.1 The Kettleman Hills oil field was discovered.
Mail-order sales were running far above the 1928 level. Department-
store sales reached a peak in September. Extra dividends were paid
in the last week of August by oil, chain-store, mail-order, steel, and flour
concerns. The agricultural situation became a matter of serious con-
cern. The United States Treasury was paying 5^g per cent in June;
federal reserve rediscount rates rose to 6 per cent by August.
Although there is a valid objection to any such statement, we may
take April to mark the peak of that (Kitchin) prosperity. But even the
inadequate description presented is sufficient to show that, whatever
may have been wrong in the financial sector, the great divisions of indus-
try and commerce either expanded2 or contracted — steel, motorcars,
building — in a perfectly orderly way during the subsequent months
through September. It is, therefore, understandable that when the
stock market — not altogether unexpectedly — collapsed, this did not
cause paralysis or even particularly strong pessimism in the business
world. What immediately happened was in fact not much more than
was foreseen, viz., a drastic reduction in the demand for "luxuries,"
of which speculative gains in stocks had been a most important feeder.
The repercussions of this were expected to induce and did induce con-
traction all round, but with money rates failing to rise to panic figures —
as compared with their reaction in prewar crises — improvement was
confidently predicted for the first half of 1930. Among characteristic
reactions of "big" business we may note that Ford announced a sub-
stantial reduction in prices, that United States Steel and American Can
declared extra dividends, and that prevailing opinion was strongly
against a decrease in wages. The withdrawal of foreign funds, the
agrarian situation, and such liquidations of concerns as occurred3 were —
quite correctly — not considered decisive.
It is of the utmost importance to realize this: given the actual facts
which it was then possible for either businessmen or economists to observe,
those diagnoses — or even the prognosis that, with the existing structure
of debt, those facts plus a drastic fall in price level would cause major
1 The following industries registered gains in employment in July : slaughtering, ice
cream, flour, pipes, structural ironwork, machine tools, furniture, shipbuilding, oil, fer-
tilizers, boxes, cement, electrical machinery, rayon, radio. Machine-tool orders reached
their peak in May. Value of exports rose to March.
2 Some of them — tin, rayon, and paper, for instance — did so vigorously in the third
quarter, a few even in the last week of the year.
3 They were — in industry and commerce — anything but sensational. In the last week
of the year, for instance, the more important ones were those of a South Carolina cotton
mill and of the Laconia Car Company. A securities company also failed during that week.
794 BUSINESS CYCLES
trouble but that nothing else would — were not simply wrong. What
nobody saw, though some people may have felt it, was that those funda-
mental data from which diagnoses and prognoses were made, were
themselves in a state of flux and that they would be swamped by the
torrents of a process of readjustment corresponding in magnitude to
the extent of the industrial revolution of the preceding 30 years. People,
for the most part, stood their ground firmly. But that ground itself
was about to give way.
F. The Behavior of Systematic Series from 1919 to 1929. — Discussion
will conveniently begin with three of the four series which enter into
our Pulse Charts (I). Then we shall discuss the behavior of (outside)
debits, incomes, and so on (II), and finally some of the series which
reflect the processes of the spheres of banking (III), speculation (IV), and
monetary management (V).
I. It should once more be emphasized that speaking of phases of an
incomplete Kondratieff involves a hypothesis which the future course of
things might easily fail to bear out, even if this future were less likely
than it is to be dominated by external factors or, if some of us resent
application of this term to government action, by public expenditure,
control, and planning. However, for the period under discussion that
hypothesis may be checked statistically, as in the preceding section it
has been checked historically, by confronting with the actual behavior of
time series the expectations which follow from that hypothesis. During
a Kondratieff recession of which the years from 1919 to (the fall of) 1925
form, according to the schema, a part, and during a Kondratieff depres-
sion to which the remaining years belong, we on the whole expect output
to increase strongly, more strongly than in the preceding Kondratieff
prosperity; price level and interest rates to fall; and balances to increase
but (barring government or central bank action and autonomous, or at
any rate external, changes in the quantity of monetary gold) less strongly
than output and also at a rate smaller than that at which they increased
in the previous upswing. Some explanations and qualifications will be
added in each case. But a difficulty which has been encountered already
must be mentioned again. Since there is no way of isolating the effects
of our process and since we can do but little more — a little more we can
do1 — than indicate the direction of the movements it tends to bring
about, we cannot numerically determine the extent to which our process
has been deflected by influences external to it, especially by influences
1 Comparison with the rates of change that obtained at comparable junctures in the
past is not, however, a method that can be trusted. We have only two such analogous
cases. And each of them displays peculiarities of its own. Even if that were not so,
there is no warrant for expecting that, for example, the rates of increase in output and of
decrease in price level should be the same in each instance.
1919-1929
795
which acted on some element of the system in the same direction as that
process itself.
a. 1. Our expectations as to the behavior of Total Physical Output
or, if we wish to exclude the influence of the short fluctuations in crops
that are due to influence of weather, of Total Output of Manufactures
(and Mining) call for the following qualifications. First, we know that
they do not apply to those subphases which we call deep depressions.
We know, in the second place, that on a Kondratieff downgrade prosperi-
ties of the shorter cycles should display greater and not smaller rates of
1897 1900 1905 1910 1915 1920 1925 1930
CHART XLII. — Industrial production (see Appendix, p. 1068).
1935
increase because, the general conditions of a Kondratieff downgrade set-
ting the stage for strong increase of output, this increase is likely to show
with particular emphasis under the influence of additional producers'
expenditure. Third, it must be remembered that for Kondratieff down-
grades in general, but especially for the one under discussion, our indices
are still more than usually likely to understate the rate of increase because
— not to insist again on the unduly great role which most indices attribute
to basic producers' goods — increased supply of intangibles (services of
all sorts), increased voluntary leisure, economies in the use of raw mate-
rials and of intermediate products, and improvements in quality con-
stitute, as we have seen, outstanding features of the period. With these
provisos in mind we will now inspect Chart XLII, which presents the
796 BUSINESS CYCLES
curves of industrial output for our three countries from the beginning of
the Kondratieff at which they have been made to coincide.1
The first thing that strikes us is the extent of the differences in both
the long-run and the short-run behavior of the three curves. Obviously
the war is only in part responsible for it. The United States curve,
keeping its more lively temperament throughout — the reader will remem-
ber that we account for this mainly by the pace of American develop-
ment— shows the Kitchins with a clearness that leaves nothing to desire.
War expenditure merely accentuated their positive phases at the expense
of the negative ones. The Juglars could not be read off correctly,
although we can trace them if we interpret the curve in the light of busi-
ness history. War effects make it difficult to test our Kondratieff
expectation. A linear trend drawn through the items from 1898 to
1912 would, however, display a smaller gradient than would a trend drawn
through the items from 1922 to 1929 which, considering that we are
using a log scale, may perhaps be held to spell acceptable verification.
But we do not insist on this. It is more significant to note that from
1899 to 1912 output of manufactures increased at most by between 70
and 80 per cent and from 1912 to 1929 by at least between 135 and 145
per cent of the 1899 figure. Figures per head of population suggest the
same conclusion. According to Professor Mills, average annual per
capita increase, excluding construction and personal services directly
consumed, was roughly 1.1 per cent for 1901 to 1913 and 2.4 for 1922 to
1929.2 Inclusion of building would, of course, strengthen these indica-
tions. Census figures of value added, corrected for price level, also
behave according to expectation.
It should be observed that there is nothing in this record of postwar
developments to support a belief in retardation and, in particular, a
belief in the restrictive tendencies that are held to be inherent in monopo-
loid large-scale business. Neither is there in it anything that could,
from experience with cyclically similar periods in the past, be called
1 Neither the curves on this chart nor the output curves in the pulse charts are strictly
comparable (see Appendix). In method as well as in quality of material and in coverage
the indices differ to a dangerous extent. Nor have the League of Nations' efforts availed
to supply us with really comparable indices. See also N. J. Wall, Monthly Index of World
Industrial Production, Preliminary Report, United States Department of Agriculture,
June 1986, and G. P. Warren and P. A. Pearson, The Physical Volume of Production in
the United States, Cornell University Agricultural Experiment Station, Memoir 144,
November 1932.
2C/. Economic Tendencies in the United States, Chap. I, especially p. 244. The
figures of average annual increase for the two periods are 3.1 and 3.8 per cent respectively,
hence also according to expectations. Mr. Leong (Journal of the American Statistical
Association, March 1932) gives 3.5 instead of 3.8. It is interesting to note that from 1899
to 1929 physical output increased by about 200 per cent. Horsepower installed in manu-
facturing plants increased by about 330 per cent.
1919-1929 797
astounding, unheard-of, or abnormal.1 In that respect the capitalist
machine seems simply to have been working as it previously did in
comparable epochs. Of course, figures of total, or total industrial, out-
put can never prove anything — either in themselves or by comparison
with money incomes — about the presence or absence of "overproduction."
Aggregates may, as we know, hide any amount of maladjustment. But
the impressive steadiness of the process2 and its perfect accordance with
the past is still important to note.
The German case hardly calls for comment. There is no need to
elaborate the reasons why war demand, impinging as it did in both
countries on the productive possibilities of a Kondratieff recession, pro-
duced the spurt in output which we should expect from this in the United
States only, whereas it caused a slump in German output. Rebound
from the trough in 1919 was first prolonged and then drastically reversed
1 The rate of increase could be made spectacular only by counting from the trough of
1921 or the beginning of 1922 to the peak of 1929, which would be devoid of sense and, in
particular, distort the comparison with England. We will notice at once that this also
applies to output per man-hour. The spectacular rate of increase in 1922 only made up
for the low ones — or even backslidings — which preceded. It has been rightly said that
industrial efficiency sank to a low level during the war and that American industry, too,
emerged from it with an antiquated and overstaffed apparatus, in fact in a drowsy state
which illustrates both an important aspect of "inflation" and the function of cold douches
such as the one applied in 1921. Measuring "efficiency" by deflated national income
divided by hours of work done, we get a minimum for 1917. For 1923 to 1929 Professor
Mills (op. cit., p. 297) gives an average annual increase in output per wage earner in 62
manufacturing industries of 2.7 per cent. For the precise meaning of this, see footnote on
that page. The increase of physical output itself during those 6 years Professor Mills
puts at 19 per cent, the average annual rate being 2.8 per cent. Of course, "secular"
shifts in the economic organism partly show in the sharp decline that occurred from 1920
to 1930 in the percentage share in the total of persons gainfully employed in agriculture,
fishing, and lumbering; in the smaller but still significant decline of the share of manufacture
and mining; and in the sharp increase of the share of trade, transportation, clerical work,
and services.
2 Again the writer largely agrees with Mr. Carl Snyder's well-known views on that
subject. But it is perhaps not superfluous to emphasize once more that an impression of
retardation such as Mr. A. F. Burns attempted to establish is perfectly understandable on
statistical and other grounds, especially in a period characterized by economies in the use
of raw materials, and that an extrapolation may easily fail to be verified by events owing
to the influence of institutional factors or of the effects on motives of the falling birth rate
or, in the end, of saturation. We will use the opportunity to quote Mr. Burns on another
point. "We may therefore conclude from our analysis of American experience since 1870:
first, that periods of sharp advance in the trend of general production, which are charac-
terized invariably by considerable divergence in production trends [quite so. J. A. S.I,
have been followed invariably by severe business depressions; second, that most of the
business depressions of marked severity have been preceded by a sharp advance in the
trend of general production and considerable divergence in the trends of individual indus-
tries." (Production Trends, p. 251.) For Kondratieff downgrades at least, we entirely
agree.
798 BUSINESS CYCLES
(1923) by inflation and the events that culminated in the Ruhr invasion.
What may be termed the Dawes recovery, which was cut short by a reac-
tion that was intensified by fiscal policy, and the upswing from the middle
of 1926 on, which, as we saw, looks so much like a Juglar prosperity,
are well marked and, very roughly, in step with American cycles. That
upswing was, as far as output figures go, relatively stronger than in this
country: the Berlin Institute's monthly index of industrial production,
for example (basis July 1924 to June 1926), reached a maximum of
129.2 in February 1928. The Wagenfiihr index of output of manu-
factures and mining, plotted on our chart, differs in details but substan-
tially conveys the same impression. But this meant little more than
that the losses of the war and inflation periods were being made up for
— unlike the United States, Germany had by 1929 gained but modestly
as compared with 1913. Output per employee increased in the aggre-
gate roughly by 13 per cent from 1925 to 1929.
As to both facts and interpretation the British case is more doubtful.
Our graph shows, of course, decline during the war — understandably
milder than in Germany — and the trough of 1921, which we interpret
in the same way as the more serious American one. We then observe
what makes a good last Kitchin of the third Juglar and the two first
Kitchins of the fourth Juglar, the beginning of the latter being blurred
by the events of 1926. But the graph also shows that the long-run
behavior of the British curve differs characteristically from that of the
two others. This difference does not date from the war but is, on the con-
trary, just as much or more marked in the first decade of the Kondraticff,
for which it has already been discussed (Chap. VIII). So far we may
accept the testimony of our graph. The post-war behavior of British
output is, however, inadequately rendered by it and has been proved,
by Professor P. Douglas, Mr. Colin Clark and other students, to come
much nearer to American achievement that the chart suggests.1 We may
confidently assume that British industrial production increased by at
least 20 per cent between the census of 1907 and 1924 and by at least
10 per cent between the latter year and 1929.2 In support and explana-
1 Both the Hoffmann index and the quarterly index of the London and Cambridge
Economic Service, which has also been plotted, are unsatisfactory as to coverage and
attribute too much weight to old and relatively declining industries. But for the prewar
period their reliability has not been called into doubt by later work, although to some
extent the same argument also applies to that time. Mr. Colin Clark (Statistical Studies
on Great Britain, Economic Journal, September 1931) accepts Mr. Rowe's index for 1907
to 1918 and calculates from this index and Trade Union unemployment percentages an
index of output per head which for 1918 only reaches the 1907 level again. To the present
writer this estimate seems too low, but there cannot be any doubt but that increase was
much below the American and German rates.
a Until 1980 a somewhat pessimistic estimate conforming to the impression conveyed
1919-1929 799
tion, the fact should be mentioned that the ratio of imported products
to total production of manufacturing industry was, throughout the
period, smaller than it had been in 1913 — industry shifting, to some
extent at least, from production for export to production for the home
market — a fact which stands out particularly clearly when figures are
corrected for those raw materials that had to be imported in both cases
(cf. Colin Clark, op. cit., p. 351). The improved picture still leaves us
with an impression of inhibited performance, but otherwise bears out
expectation as to the behavior of industrial output in a Kondratieff
downgrade.
2. Again, there is nothing suggestive of abnormalities or new
problems in the behavior, relatively to each other, of the aggregates
into which output of manufactures is usually divided. In order to
exemplify this, it will suffice to observe the behavior in this country of
producers' and consumers* goods, on the one hand, and of durable and
transient goods, on the other (Chart XLITI).
by our chart very generally prevailed, in spite of certain objections that were raised against
it (as, for instance, by Mr. G. L. Schwartz, Economic Journal, March 1929). Professor
Bowley and Sir J. Stamp (The National Income 1924, 1927, p. 55) arrived at the conclusion
"that real home-produced income per head (of population) was very nearly the same in
1911 and 1924; it is improbable that it was any greater in the latter year, and it may have
been 4 per cent less." And the London and Cambridge Economic Service Bulletin for April
1930, stated that from 1924 to 1928 output per head (employed) was " stationary or
decreasing." This was, on the strength of new evidence, corrected in the Bulletin for
June 1930 so as to read that there had been during those years an increase of 4 per cent,
and another of 7 per cent from 1928 to 1929. Mr. Colin Clark (op. cit., p. 857) comparing
the value figures of the 1907 and 1924 census of the Final Product of Industry minus
Primary Materials Purchased by means of a more appropriate price index arrived at a
Net Output of Industry of respectively 675 and 820 million pounds of 1907 purchasing
power, an increase of 21.5 per cent. Douglas and Tolles (Journal of Political Economy,
February 1930) proceeding by another route even arrived at 23.5 per cent. As pointed
out by Mr. Clark, it would follow that real output per worker increased by about 10 per
cent, in spite of the reduction of hours. These estimates, so it seems to the present writer,
cannot be far from the truth although, for the purposes of a comparison with the United
States, it must not be forgotten that downward bias is not altogether absent from American
indices either. For such a comparison see A. W. Flux, Industrial Productivity in Great
Britain and the United States, Quarterly Journal of Economics, November 1933. For
the behavior of industrial output from 1924 to 1929 both the Board of Trade Index of
Production and the Annual Index of the London and Cambridge Economic Service — the
latter after elimination of agriculture and dwelling-house construction which it contains —
may be alternatively used. Using the former and deducting from the figure of all insured
persons in employment those engaged in building, transport, distribution, and services,
Mr. Clark (op. cit., p. 360) calculated another increase of output per head in manufacturing
and mining of 10 per cent for 1924 to 1929. Mr. E. Devons (Economic Journal, September
1935) gives 15.1 per cent after excluding from the employment figures all items not covered
by the Board of Trade index. On the London and Cambridge Economic Service indices
see also Rowe, An Index of the Physical Volume of Production, Economic Journal, June
1927.
800
BUSINESS CYCLES
We find the divergences which we should expect and, in particular,
we see that the output of consumers' goods as a rule (notably in 1924 and
from the beginning of 1930 on) decreases less strongly in troughs and
increases less strongly on upgrades than does the output of producers'
goods. But we also find, exactly as we did in the case of comparable
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
CHART XL1IL — United States production series (see Appendix, p. 1068).
prewar series, that the difference is, excepting the years from 1919 to 1921
and those of the world crisis, neither so strong nor so regular as some
theories suggest. On the whole, both series move well together.1 Again,
1 This holds for other countries also. See, for example, Statistical Yearbook of the
League of Nations, 1931, Table 91. In Germany output of producers' goods (1928 = 100)
was 97 in 1927, 106 in 1929 and 82 in 1930, output of consumers' goods (1928 = 100)
respectively 99, 94, and 86. The fact that the latter should have fallen so considerably
while the former was still rising is partly due to the particularly strong downward bias
of that consumers' good index (textiles, motorcars, boots, china, pianos, watches and
1919-1929 801
the really significant difference is between durable and nondurable goods.
Since the relative importance in the consumer's budget of the former
naturally increases with increasing wealth, we have here a factor — though
not necessarily a decisive one — which makes for increasing amplitudes of
cyclical fluctuations: to take an outstanding instance, new passenger
car registrations dropped from about 3.8 millions in 1929 to 1.1 millions
in 1932. 1 It is interesting to note that while production of durable
producers' goods and production of durable consumers' goods moved
substantially parallel,2 yet the latter increased more strongly. The
average annual rate of increase in the production of industrial equipment
— including nonresidential construction — was, according to Professor
Mills' estimate, 6.4 per cent for 1922 to 1929.
Of course, excepting special cases, we do not know with any exactness
how much of this was mere replacement, how much was replacement
plus improvement, how much was net addition, and how much of that
served new purposes. Even if we did, comparison of this figure with
prewar rates might easily be misleading, because the period under dis-
cussion happens to end with a Juglar prosperity and a Juglar recession
in which real investment would be higher than usual. But Professor
Kuznets' figures suggest that real investment was much more modest
than is very generally believed3 and lend little support to any of the
current theories of the overinvestment type. This, however, does not
mean that production of equipment goods had nothing to do with the
crisis, for it undoubtedly reflected the process of industrial reconstruc-
tion, which was going on at an increasing rate that implied quick obso-
paper), yet is not wholly spurious. Investment in equipment industries was an outstanding
feature during the years from 1926 to 1929.
1 Figures from R. L. Polk and Co., as published in Automobile Facts and Figures, 1937
ed., by the Automobile Manufacturers' Association. There was, moreover, the shift
toward cheaper cars.
2 Together, durable-goods industries of both types employed, in the average of 1921 to
1930, about 52 per cent of all factory workers (estimate by Colonel Ayres).
3 See S. Kuznets, National Income and Capital Formation, 1919-1935 (1938), Table 13,
p. 48. The relevant item is II, 1, 6, net capital formation by business exclusive of net
changes in inventories (which are of a special nature and should not be lumped together with
plant and equipment; the reader will furthermore recall the reasons why we exclude
residential building and "public capital formation," excepting productive establishments
owned by public bodies). At 1929 prices the average of 1919 to 1929 is only 2.5 billions.
But owing to the course of the cyclical phases, the series moves successively on three
different levels: around an average of 1 billion, 1919 to 1922; then for 2 years at 2.3 billions
per year, while for 1925 to 1929 the average is 3.7 billions. This also shows (the 1929
figure, 4.3 billions, is the highest) that there was no slackening in real investment previous
to and independent of the crisis itself. The employment index for the machine-tool
industry (1926 = 100) gives the same impression. It was 82 for 1924, 85.8 for 1925, 100
for 1926, 92.8 for 1927, 100,8 for 1928, and 129.8 for 1929.
802 BUSINESS CYCLES
lescence and was bound to enforce drastic adjustments and to pass death
sentence on many plants and concerns. That process is, as we know,
the very soul of every recession or depression. Some of these adjust-
ments, in the motorcar industry for instance, were carried out currently.
Others were not, and from this source flowed what may indeed be
called an accumulation of maladjustments, though it would be wrong to
attribute it to an overproduction of equipment. Extensive liquidation
became necessary — not, however, in order to reestablish an equilibrium
disturbed by overexpansion in aggregates, but in order to reestablish an
equilibrium disturbed by innovation the roots of which dated back to
the nineties. This, of course, formulates the matter in terms of the long
wave only, but the influence of the Juglars and Kitchins can easily be
inserted.
Other evidence strengthens the case for interpretation in our sense.
Among other things, the attitude of businessmen accords very well with
it. As pointed out in the preceding section, we find throughout the
twenties and in the midst of prosperities businessmen struggling against
the results of their own actions — those phenomena repeatedly noticed
in this book (for example in Chap. Ill) which, ranging from talks about
overproduction and lack of purchasing power to support given to cer-
tain types of corporative action and planning, were eventually to find
expression in the NRA legislation, the idea of which characteristically
enough originated in business circles. This is precisely what we should
expect to observe in the course of the competing-down process. It is a
question of some interest whether these tendencies were any weaker in the
twenties and seventies of the nineteenth century. Certainly, they met
with less success than they eventually did in our own time. Capitalist
governments and parliaments were much less anxious to guarantee
profits and to protect from losses than their less capitalist-minded
successors. And there is really no paradox in this.
Moreover, throughout the twenties and even in the heyday of pros-
perity phases and high profits there was considerable excess capacity1
1 It has been pointed out before that excess capacity in any relevant sense is extremely
difficult to establish and to measure, and why this is so. The work by Nourse and Asso-
ciates, America's Capacity to Produce, has been severely criticized. But it seems per-
missible to say that the evidence there presented suffices to support the statement in the
text. The same may be said of the study of a number of industries by R. F. Martin, Bul-
letin of the Taylor Society, June 1932: nearly every one of the industries covered displayed
underutilization of capacity in every year, 1921 to 1932. The question was also investi-
gated by a subcommittee of the Columbia Commission on Economic Reconstruction (see
Report, Appendix I). First, that committee was permitted to use the results of a study
by L. P. Alford and J. E. Hannum on four industries, and second, it issued a questionnaire
of its own to a considerable number of engineers and executives. Results are more relevant
to efficiency (as measured by product per man-hour) than to degree of utilization and give
1919-1929 803
which was not confined to ailing industries or regions nor entirely explic-
able on oligopolistic strategy. The reader knows how many funda-
mentally different situations are covered by that term and it is clear
that instances of all of them could easily be adduced. The one that is
relevant here consists, first, in the presence of old plant — it need not be
old in years — simultaneously with new plant and, second, in the fre-
quent necessity of erecting plant and equipment on a scale intended to
provide for future developments of demand. As we have seen before,
the more progressive an economic system is, the more of these types of
excess capacity it must display — and, also, would have to display in a
socialist state. This has palpably been the case in this country during
the twenties. There are few industries, such as oil and shoes, in which
presence of plant not used up to economic optimum is clearly of another
type, i.e., not incident to some process of change. The usual reasoning
about excess capacity in general, therefore, entirely misses the salient
point, but the relevance of the facts so designated to the explanation of
the subsequent crisis becomes only the more obvious. Planned resistance
by producers is, of course, likely to accentuate consequences.1
3. Finally, like excess capacity, supernormal unemployment is — pre-
cisely in association with strong increase in output — an essential element
of our picture of Kondraticff downgrades. We have observed it — each
time in that characteristic association — in the two previous instances,
and we find it in this. The postwar unemployment is, however, a com-
plex phenomenon and will have to be touched upon again below. Here
it is only the technological or rationalization component which calls for
notice. Some facts and figures about product per man-hour or wage
earner have been mentioned already, which also show once more how
misleading must be any theoretical schema that assumes proportionality
between output and employment for any period longer than a Kitchin.
In further illustration we will glance at a few data taken from the investi-
gation of the Berlin Institute on unemployment in Germany.2 Product
per workman increased from 1926 to 1929 (1930) by 25 per cent in the
rise to various doubts. It is, however, interesting to note how (relatively) often the
question what factors had been responsible for keeping output below the — widely differing
—estimates of possible maxima, has been answered by "competition and substitution."
As might be expected, lack of effective demand was the most frequent answer and the one
which seems particularly to have impressed the committee. In view of the inference
drawn from it, it is perhaps not superfluous to state that it does not follow that inadequate
money incomes were the villains of the piece.
1 "For we think that the violence of the convulsions such as recurred 1907-1908,
1920-1921, 1929-1933 is due largely to the partial character of the liquidation effected
during mild contraction." Professor W. C. Mitchell in National Bureau of Economic
Research, Bulletin 61, for Nov. 9, 1936.
2 A. Reithmger, Stand und Ursachen der Arbeitslosigkeit in Deutschland, 1932.
804 BUSINESS CYCLES
iron, steel, metal, machine, and motorcar industries; by 18 per cent in
mining; by 16 per cent in the woodworking industries; by 15 per cent in
building; by 13 per cent in the chemical industry; by 10 per cent in the
paper, printing, food, and textile industries; by 5 per cent in the leather
industry. Comparable figures of employment are not available in all
cases. But we know, for instance, that in the production of crude steel
employment fell between 1925 and 1927 by about 10 per cent; in coal
mining between 1925 and 1929 by over 5 per cent, in the motorcar indus-
try also by about 5 per cent; while in the machine, paper, printing, elec-
trical, chemical, building, food, textile and clothing, woodworking, and
ceramic industries there was (mostly modest) increase. Fuller data for
output per wage earner or even per man-hour in individual industries are
available for this country. Reference should again be made to Professor
Mills's Economic Tendencies.1
But while increase in product per man-hour may be used to illustrate
one aspect of that period of rationalization, it is not superfluous to repeat
that — irrespective even of purely statistical questions that may make
all the difference in results — it must for the purpose in hand be used
with more caution than is usually applied. It does not, of course, either
isolate or measure the technological component, if indeed the word
component be permissible at all, of total unemployment or the effect of
innovation on total output. It is a gross error to look upon change in
product per man-hour and change in total production as two independent
factors militating against each other and producing actual employment
as their resultant. Moreover, nothing about "secular" trends or tend-
encies inherent to the system can be inferred from the survey of the
facts of so short a period, in which ultimate results cannot possibly have
had time to show their true faces and which was one of rapid readjust-
ments. Nevertheless, we may confront the impression conveyed by
figures about man-hour product with the behavior of employment on the
understanding that we are dealing with but two of many interdependent
variables and that no conclusions about causal relations can be derived
in this way.
Total employment increased, of course, in all three countries. Part-
time employment — to which it would strictly be necessary to add uneco-
nomic employment, especially in agriculture — changes in school age and
in enforcement of attendance at school, in the employment of women, in
age distribution, for Germany also the abolishment of compulsory mili-
tary service and partial annihilation of the rentier class not only make
comparison with changes in population more difficult, but also deprive
it of much of its meaning. But it may be stated that available indica-
tions for this country and for Great Britain point to the conclusion that
1 See in particular pp. 296 and 297.
1919-1929 805
total employment increased from the end of 1922 to the end of 1929 in
such a way as to produce no — or, if the reader prefer, a substantially hor-
izontal— "descriptive trend" in unemployment percentage. Although
that level was, as we should expect, considerably higher than during the
Kondratieff prosperity, there must, hence, have been "absorption of the
technologically unemployed" and of the increase in employable popula-
tion within that period in this sense, although in no other.1
Of course, this does not apply to any individual industry or group of
industries. Manufacturing, or manufacturing and mining, forms, for the
purpose in hand, an arbitrary group. Also, since it includes the most
"progressive" branches of economic activity, it is not a random sample.
Hence, no conclusions can be drawn from its conditions alone. In all
three countries employment in manufacturing industries increased but
slightly, if at all. In Great Britain the number of persons employed in
manufacturing and mining increased from a little below 6 millions in 1924
to a little over 6 millions in 1929, while the corresponding number for
building, transportation, distribution, and services increased by 600,000.
In Germany industrial employment according to the Institut fur Kon-
junkturforschung increased from 1925 to the middle of 1929 by about
^ million, or 5 per cent.2 In this country census figures reflect the great
increase that occurred during the war period in the number of wage
earners employed by the manufacturing industries covered. This
increase was practically wiped out between 1919 and 1921, but nine-
tenths of the loss of those years was recovered from 1921 to 1923. The
next two biennial periods showed a loss of about 470,000, the period
from 1927 to 1929 a gain of about 360,000. Comparison via census
pay-roll figures gives between 1919 and 1929 a loss of about 160,000, while
by extending coverage the loss from 1923 to 1929 can be converted into a
small gain.3 Over the period there was thus substantial constancy
1 No such statement can be made about Germany because of the — "stabilization" —
peak in unemployment which occurred in the winter, 1923-1924. If we exclude it, then
the "trend" will depend on our choice of a precise spot between that peak and the minimum
of the summer 1925. From the minimum in 1927 to the end of 1929 the gradient of the
"trend" would be strongly positive, but that does not mean much especially as it may be
accounted for by circumstances not directly relevant to our present subject.
2 Reduced to comparability, however, that figure would, the writer believes, dwindle
to still more modest proportions.
8 C. Goodrich and Associates, Migrations and Economic Opportunities, 1936, puts the
increase from 1928 to 1929 at 185,000. Professor Mills, op. cit., p. 480, stated that
employment in all manufacturing industries increased at the rate of 0.6 per cent a year
between 1922 and 1929. But this is as likely to be misunderstood as the statement that
employment "declined" from 1923 to 1929 (see Aspects of Manufacturing Operations,
National Bureau Bulletin, May 1935). Mr. D. Weintraub (Displacement of Workers
by Increase in Efficiency and their Absorption by Industry, 1920-1931, Journal of the
American Statistical Association, December 1932) found that no "permanent" displacement
806 BUSINESS CYCLES
partially reflected in the drastic fall in labor cost per unit of product
in the manufacturing industries from 1921 on (see Mills, op. cit.t
p. 413.) This does not prove, of course, that developments in manu-
facturing did nothing toward providing employment for the increasing
population, because the sources of employment which took care of the
increment in labor supply were directly and indirectly created by it.
Nor does it prove that the unemployment of the period can be exclusively
described in terms of the technological factor. l But it does prove that the
latter actually did play the role which we attribute to it in pur picture
of Kondratieff downgrades.
can be proved for the period 1920-1926, greater numbers having been absorbed "through**
increase of output than were displaced "by" increase in "efficiency," while 1926-1929 only
about 0.1 per cent of the workmen displaced failed to be absorbed, but that, of course,
temporary displacement was considerable — about % million per year. See, however,
J. Lubin, Absorption of the Unemployed by American Industry, Brookings Institution,
1929. On the whole, those results do not seem to diverge from either our impression as
to the facts or our interpretation. Much information that is, in spite of doubts on statis-
tical and theoretical grounds, very valuable has been forthcoming of late as a result of a
National Research Project of the WPA. See, in particular, D. Weintraub and H. L.
Posner, Unemployment and Increasing Productivity, March 1937. We will mention
only a few figures about the system as a whole, without attempting to appraise their precise
significance. Taking 1920 for base, the authors give 146 as index of total physical national
product for 1929, 126 as index of product per man-year, and 118 as index of total available
(usable) labor supply. It would follow that during that period the system as a whole
"absorbed" much more than the technological unemployment it "created," in fact, almost
that unemployment plus the simultaneous increase in available supply of labor — a result in
substantial accord with such other evidence as we have. Such statements are, however,
open to a well-nigh innumerable host of objections, and certain obvious inferences cannot
fail to rouse the ire of the more ardent critics and defenders of capitalism. The ones will
object that for many workmen what looks statistically to be temporary loss of employment
often is permanent, that reemployment is very often secured after delay spelling much
hardship, that even when secured the new job may be transitory, less skilled, or otherwise
less desirable, that the "services • • • are merely a buffer margin to enable the present sys-
tem to frustrate its own genius in the interest of its creditors" (Mr. A. McLeish, Machines
and the Future, The Nation, Feb. 8, 1938), and so on. The others will point to the steady
and considerable rise in total money and real pay rolls which is relevant to the question,
not only sub specie of a "compensation" to the working class as a whole, but also because
it does not necessarily follow that displacement would have been as widespread if wages
had increased less (see below); to long-run tendencies such as are displayed by the net
increase in jobs since 1900; to the arbitrariness of a standard which takes the superiority
of alternatives for granted, and so on. Any social system stands a priori condemned by
some and a priori justified by others. But there is not much room for disagreement
about the facts.
1 No unemployment ever can, except in the shortest of runs. When Professor J. M.
Clark (Economics of Planning Public Works, National Planning Board, 1935) testified
to his belief that "mere technological progress seems capable • • • of bringing on a state
of chronic inability to use all our labor power,'* he wisely inserted "lacking the necessary
adjustments." With this, to be sure, everybody will agree.
1919-1929 807
b. Next, turning to price levels in order to see how far the facts lend
themselves to interpretation in the light of our model,1 we must first
qualify our expectation for Kondratieff downgrades by taking account of
the influence of Juglar and Kitchin phases. Then we must bear in mind
that price indices will understate the real decline, for the same reasons
why quantity indices understate the real increase plus the additional
one, that in some cases the prices actually paid are lower than the quota-
tions which enter into the indices. Finally, we have to add the second
great factor that made for decline, the reaction of the system to the war
disturbance.
What we have before us is obviously the combined effect of both.
Even in theory, let alone numerical operation, it would be extremely dif-
ficult, if not impossible, to divide it up between them. The two following
sets of propositions are worth stating, however.
First, if the war disturbance had impinged on an otherwise stationary
process and if no permanent change had occurred in the monetary sphere
— the war being financed merely by straining an existing and unchanged
monetary and credit apparatus — then prices would eventually, in the
course of a process of repayment of government debts from taxation, have
fallen to prewar levels. Any permanent expansion of the sphere of
money and credit, especially if carried out by all countries in step, would
pro tanto have removed this mechanical "deflationary" pressure. But
this does not mean that, in the absence of it, no fall at all would have
occurred, for the mere process of normalization in the business sphere, the
cessation of war demand, the reopening of blocked channels of trade, the
resumption of normal production and habits would have been sufficient
to cause both a downward jerk and a permanent lowering of price levels
from war peaks: the self-deflation of business cannot be prevented by
mere abstention from "deflation," but only by continuing "inflation"
by additional government expenditure.
Second, if there had been no war and if no autonomous change had
occurred in the sphere of money and credit, then it follows from our model
that by about 1925, when a neighborhood occurred for all three cycles,
the price level should have been somewhat below that of the preceding
three-cycle neighborhood, viz., of 1897. From about 1911 the steady
1 In a sense this may also be called an attempt to delineate, and to compare with the
actual, the " natural" course of events. But, if we chose this mode of expression, we should
have to recall once more our distinction between equilibrium influences and equilibrating
influences. A priori there need not be more virtue in a "natural" economic process than
there is in the natural course of tuberculosis. And all that we actually claim for the former
is that it fills certain "physiological functions" — a turn of phrase which need not again be
justified — and cannot in genera) be interfered with without producing as well as removing
undesired results. In this respect there is, however, a great difference between the reaction
of the system to the war disturbance and its reacting to our process.
808 BUSINESS CYCLES
pressure of " progress" — no need to explain this again — would have
increased real incomes by enforcing a steady fall, which as we know
should, in spite of so much permanent expansion of the monetary sphere
as our process would of itself entail (see Chaps. IV, XI, and XII), have
landed the price level at some figure below that of the neighborhood from
which the current Kondratieff started. If it did not, this must have been
due to the war disturbance and other factors of monetary expansion, such
as the increase in gold stock since the nineties or the increase of lending
facilities through banking legislation, both taking effect in successive
prosperities. But autonomous change in the sphere of money and credit
does not per se eliminate the systematic tendency of price level to fall in
a Kondratieff downgrade. It only raises the level on which this tendency
acts. Hence, if, say, devaluation had been resorted to in order to pre-
vent the price fall of 1920 and if it had been successful in preventing it, the
period would still have displayed that tendency and the phenomenon
which for us is perfectly natural but for many economists a paradoxon,
viz., prosperity with falling prices. This fall we would expect to be
interrupted, or temporarily turned into a small increase, around the
beginning of the fourth Juglar, and then to resume at an accelerating
rate as depression approached.
1. Inspecting the curve traced out by the price level in this country
as represented by the B.L.S. index of wholesale prices (Chart XXXIX),
we first of all notice the drop from the (monthly) maximum of 167.2
(1926 = 100) which occurred in May 1920, to a minimum of 91.4 in
January 1922. This drop, or rather, if we allow for the fact that panic
declines always outrun the goal and that hence some rebound would have
occurred even in the absence of any other influence, a drop to a few points
above that figure1 was due not merely to normalizing reaction to war
disturbance: we know that this reaction coincides with a Juglar depres-
sion. But that reaction was, it is safe to say, the dominant factor in it.
We also know and shall definitely see presently that no appreciable
"deflationary" pressure was exerted on the system. The huge machine
for credit creation set up by the Federal Reserve Act as amended during
the war period was left intact, and war expansion of the credit structure
was supported and to some extent camouflaged by the broadening of
its gold basis. As a postwar adjustment that drop was, hence, almost
wholly due to the self deflation of business.
Now, that minimum of 91.4 was still about 30 per cent above the
annual average of 1913, and it is quite possible that reaction to war
1 This allowance is a difficult matter, for what it is reasonable to attribute to rebound
from a trough exaggerated by the spiral, is mixed up with the effects of Kitchin movements
and other influences and cannot simply be read off either from the maximum attained in
March 1923 (104.5) or from the broader plateau of about 99 to which price level then
relapsed.
1919-1929 809
inflation continued to play some role throughout the twenties. But
owing to the circumstances just mentioned, this effect cannot have been
significant. At least it is as safe to say that the rest of the period was
dominated by other factors as it was to say that 1920 to 1922 was dom-
inated by that reaction, and it is on this rest of the period that we have to
try out our schema. Rebound from panic low plus the advent of Juglar
recovery plus Kitchin effect may then be invoked in explanation of the
rise in the B.L.S. all-commodity index to March 1923, which was mainly
due to the class of semimanufactured commodities (128.3 in April 19231).
It was suspiciously strong and, in fact, gave way to relapse after the first
quarter of 1923 (annual figure for the year — monthly average — 100.6).
During 1924 there was substantial stability at the lower level reached in
the last three quarters of 1923 (annual figure, 98.1), in 1925 a rise (annual
figure, 103.5) followed by a fall during 1926 (97.9 for December), tapering
off in 1927, interrupted in 1928, resuming in the last quarter of that year,
and gaining momentum in the last quarter of 1929 (December: 93.3).
Finished commodities fell practically steadily from the maximum of
102.1 (November 1925). 2 The index of sensitive prices came down well
in 1926, rose somewhat in 1927, sagged in 1928, and slumped in the last
quarter of 1929. Comparison of the "domestic" price level with those
of exports and imports3 would greatly add to the details of the picture
without altering its general contours. The National Industrial Con-
ference Board's index of cost of living falls from nearly 200 per cent of its
1914 base (the middle of 1920) to a little above 160 (the middle of 1922),
rises to the middle of 1925, and then falls slowly, not, however, reaching
160 before the middle of 1930.
This is what has been so frequently referred to as the stability of the
American price level4 during the twenties and taken as proof of the
1 This fact, however, would accord well with the "rebound theory," since the prices
in that class display the highest of the maxima in 1920 (258 in May) and the sharpest fall
(90.3 in December 1921).
2 If we eliminate farm products, the picture is from our standpoint greatly improved :
we find steady fall from 1923 on at an average rate of 1.5 per cent per year, varying, how-
ever, nicely in the phases of the two shorter cycles (see Professor Mills, op. cit.t Fig. 61 on
p. 341). The same applies to processed producers' goods and from 1925 to processed
consumers* goods, but neither to raw producers' nor to raw consumers' goods (see ibid.,
Fig. 68 on p. 359).
3 See, for example, Professor Theodore J. Kreps, Export, Import and Domestic Prices
in the United States, 1926-1930, Quarterly Journal of Economics, February 1932. One
point is, of course, obvious, viz., that export prices of manufactured goods were in general
lower than the domestic prices of the same goods and behaved, in their short-run tendencies,
more according to the competitive pattern.
* It must be remembered not only that it was the fashion to exaggerate such stability
as there was but also that there was a time, bygone forever let us hope, when many econ-
omists believed that keeping the price level constant is all that is necessary in order to
insure general stability of the economic system, to avoid depressions, etc.
810 BUSINESS CYCLES
success of federal reserve policy and/or of the absence of "inflation."
We shall return to this later on. Meanwhile, it is submitted that the
price behavior described does not fail to bear out our expectations.
The Kitchins show well, the Juglar effect is not absent, and the Kondra-
tieff pressure due to the long-term effects of "progress" is much in
evidence. In the light of later events and considering the extent of the
expansion in physical product and of the fall in real cost, an impression
may dawn upon us to the effect that the slightly inclined plateau we
behold — from 1922 to 1929 the rate was about % per cent a year — may
perchance not have been all that we should have expected from the
unhampered working of our model, especially when we weigh the inter-
ruption and partial reversal of the falling tendency from the summer of
1927 to the summer of 1929. And from this we might go on to infer
that there was some factor which prevented prices from falling as much as
they would have if left to themselves. But so long as the presence of
such a factor is not proved, we have not more than an impression which
may easily mislead.
2. The English case, fundamentally similar in other respects, differs
from the American by the element of significant pressure exerted by the
monetary factor, which was present in the former and not in the latter.
And since under English conditions this pressure would primarily work
through affecting imports and exports, devaluation of the pound in 1922,
when the price level had had the opportunity of showing its true face, to
62.5 per cent of the prewar par, the figure roughly corresponding to that
of the Board of Trade price index for that year, would in this particular
case, as has been pointed out in Sec. C, have in fact removed a weight
that not only depressed but also distorted. The Board of Trade index of
wholesale prices displayed the postwar drop, falling from 325 per cent of
the 1913 annual figure (April and May 1920) steadily through 1922
(December: 156). Then it rose no less steadily through 1923 and 1924
(maximum of 171 was reached in January 1925). That both drop and
ascent lasted longer than in this country is not difficult to understand.
But "overvaluation" of the pound must be invoked in explanation of
the subsequent fall to 144 in April 1926. The disturbances of that year
make it unsafe to speak of an effect of a Juglar prosperity being reflected
in the modest rise to November (152), after which the English level,
falling almost steadily to 136 for September 1929, conforms to expectation
better than the American one.1 The Ministry of Labour cost of living
index (October 1920, 276 per cent of 1913; December 1929, 166) inter-
rupted an otherwise steady downward course only in 1923—1924.
1 But the extreme stability, at about 80 per cent of the 1924 level, of the Board of
Trade's index exclusive of foods during 1927, 1928, and three quarters of 1929, somewhat
militates against that statement.
1919-1929 811
3. The German case, finally, is almost completely dominated by fac-
tors extraneous to our process. The annual index of the Institut fiir
Konjunkturforschung (1913 == 100) resumes in 1924, the first year after
the gap caused by the war and by "wild" inflation, at 136. Prices of
agrarian products were then at 112, sugar at 176, products from over-
seas at 125, and industrial materials at 146. 1 These figures obviously
reflect past disturbance, both in their absolute values and in their rela-
tions to each other — the new prices in terms of gold marks being largely
experimental — and the rise of the index in the next year and its drop to
128 for 1926 was but the result of adjustments which have little to do
with normal cyclical movements. From the middle of 1926 to the
middle of 1928 the general movement is upward and thus contrary to
the English and American tendencies, though shorter fluctuations arc
fairly well in step. This may, in part, have been due to the influence of
the prosperity phase of the fourth Juglar, but much more important
was that of the "consumers' prosperity." The influx of the foreign
credits that helped to finance both was not, of course, an additional factor
but only implemented the two others. Since, however, the supply of
those credits constituted an independent element of the situation, a
surface explanation can alternatively be presented in terms of them.
This will account both for the rise in prices in 1925 and for their trough
in 1926, but it will not account for the tapering off in 1928 when beginning
weakness of prices preceded the withdrawals of foreign funds. We may,
moreover, note the heavy responsibility of building materials for the
rise of the index in 1927 and 1928 and the insignificance of the contribu-
tion of nonferrous metals and of chemicals — facts which tend to reduce
the extent of the abnormality. The cost-of-living index rose practically
without interruption from the beginning of 1924 to March 1929, and
thus presents the problem implied in behavior contrary to expectation
still more clearly than the index of wholesale prices. But explanation
by the agricultural policy and by the "consumers' prosperity" also
becomes still more convincing when applied to cost of living.
c. The attempt to interpret experimentally, in the same way as
price levels, the behavior of interest rates in the light of war disturbances
and of our model is still more hazardous, because the influence of
"policies" and of other external factors — open-market operations in the
United States, the dominant role of government financing and of the
management of the pound in England, all sorts of irregularities in Ger-
many— may still more obviously than in the case of price levels seem to
preclude all reasonable hope of finding traces of our process. This is a
question of fact, however, and at the outset it is sufficient to answer
1 Since practically no finished products enter into that index, we cannot trust it to
render the real behavior of the price level faithfully.
812 BUSINESS CYCLES
doubts on that score by pointing out that policies and other external
factors manage or influence money or credit not in abstracto but in given
business situations which, barring the theory that they are nothing but
the products of monetary policy, must assert themselves both in the
measures taken and in their effects.
Expectation is for a fall, possibly interrupted by Juglar and Kitchin
prosperities. This applies to both the war and the cyclical component.
We shall, as in the case of price levels, attribute partly, though not
wholly, the peak in 1920 as well as the fall in 1921 — which in this country
preceded open-market purchases and may, hence, be looked upon as
"genuine," while the additional fall in the first half of 1922 raises another
problem — to the war and its liquidation, and apply expectations from
our model primarily to the rest of the period, although, especially for
England and Germany, conditions traceable to the war continued to influ-
ence rates still more obviously than price levels. It should be observed
at once that our theoretical expectation being what it is, a Hayek effect1
of cheap money policies must be extremely difficult to discover and pre-
sumably cannot be discovered at all from the behavior of interest rates
alone.
1. The strongest traces of what, from the standpoint of our model,
would be normal behavior we shall, of course, expect to find in the United
States. Although the advent of the Federal Reserve System materially
changed the significance of the commercial paper rate,2 we may still use
it to represent the course of short rates, especially as acceptance, call,
and time money rates would not give significantly different results. If
we take our stand at the end of 1921, when it was at about 5 per cent and
the postwar hump had been a little more than eliminated, we do in fact
find the falling tendency we look for until the end of 1927 (see Chart
XXXIX). The trough in the second half of 1924 strikes us as abnormally
deep, but the increase in 1925 at the beginning of the fourth Juglar and
the substantially level movement during the two following years are
exactly what we would expect, especially if taken together with the fall
that occurred in 1930. But between there is another hump: the rise in
1928 and to nearly the middle of 1929 is distinctly contrary to expecta-
tion and calls for an explaining factor, which we shall have no difficulty in
discovering. An obvious clue is, however, afforded by the leadership in
that movement of call rate. The Kitchins show well.
1 By a Hayek effect we simply mean the effect on investment of a rate of interest lower
than would have obtained had the process been left to itself. We are not discussing the
applicability of Professor Von Hayek's theory in its strict acceptance, the assumptions of
which do not seem to be fulfilled by postwar conditions.
2 The behavior of commercial paper rate in fact differs from the prewar pattern. The
seasonal movement, for instance, has all but disappeared.
1919-1929 813
Bond yields1 illustrate our point still better (see Chart LIT). From
the end of 1921 to the first quarter of 1928 they (Moody's index of AAA
bonds) declined by a little less than 1 per cent and this fall was, barring
the recovery from the trough of 1922, almost without interruption.
Their increase from the second quarter of 1928 to the third quarter of
1929 was but small and is easily accounted for by the abnormal behavior
of short rates and by the speculative mania in general. Though the
relation over the whole period between short rates and bond yields
differs from their prewar relation, the latter thus still give the "trend" or
general drift of the former. An impression to the contrary is created
merely by the 1924 trough in short rates not reproduced in long rates
and yields. Since the rise in 1928 and half of 1929 was also but very
weakly reproduced in yields, formal trends drawn through money rates
and yields from the middle of 1924 to the middle of 1929 would, of course,
show divergent movements. But there is no justification for such a
procedure.
It has been pointed out in Chap. XII that the rate on customers'
loans is not in practice what it is in the theoretical blueprint, viz., the
core of the interest structure. In Germany, in particular, it followed
mechanically the Reichsbank rate, and fixing it seems to have been
almost a clerical job. But in this country it still retained something of
that pivotal significance which our schema assigns to it. Therefore, it
seems worth our while to glance at its variations (see Chart XLIV).
1 Reference is again due to Mr. C. C. Abbott's important studies, Review of Economic
Statistics, January and May 1935. It is a matter of regret to the present writer that this
book had to go to press without the benefit it would have derived from the researches of
Mr. F. R. Macaulay, Movements of Interest Rates, Bond Yields and Stock Prices in the
United States since 1856, 1988. By familiar operations, which are very simple in the case
of the United States but require "corrections" involving the exercise of much arbitrary
judgment in the other two cases, a rough covariation between the short-run variations—-
mainly the Kitchins — in interest rate and in price level could be brought out graphically.
We do not do this because all that matters for the purposes of this book is evident from
inspection of the pulse charts. It will also be evident that even in the American case the
relation is clearly a disturbed one — which must be so quite apart from external disturbances,
since rates and levels are not the only elements of even the aggregative system — but it is
fairly close from 1923 to 1927. If the opposite "underlying tendencies" are eliminated,
rough covariation is still more obvious between interest and industrial output. In a paper
previously quoted (Journal of the American Statistical Association, June 1931, p. 5) Mr.
B. B. Smith compared the (smoothed) business index of the Cleveland Trust Company
with the (smoothed) deviations of a short money index from bond yields (the latter taken
as "normal," an idea which corresponds well with our view of the matter) lagged by one
year, and found an acceptable inverse covariation between the two through 1928. A little
reflection should convince the reader that, although we could not stand for a lag of exactly
one year — but only for a shorter one — this result lends itself to interpretation by the
Kitchin phases. It should be observed that it does not contradict the "Harvard lag" of
short rates.
814
BUSINESS CYCLES
An average for 26 to 35 cities computed by the National Bureau of
Economic Research (Bulletin for July 18, 1927) moved much like bond
yield, though more sluggishly and at a level that was considerably — by
about from % per cent to 1 per cent — higher. Only in New York City
rates on customers' loans moved until the summer of 1927 at about the
same level as bond yields, reflecting as they should the rise of the fourth
Juglar better than these, but responding more strongly to the abnormal
conditions that obtained from the middle of 1927 to the middle of 1929.
1920 I92f 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
CHART XLIV. — United States; rates on customers' loans (see Appendix, p. 1069).
In the South and West and even in the North and East those rates
throughout the period stayed at a figure much above Kondratieff expecta-
tion, and if it were not for the fact that other and much cheaper sources
of money were readily available to all that was most "progressive'*
in the business organism of those years, we should be faced with the
question, not of the presence of a Hayek effect, but, on the contrary, of
the effects of dear money.1 As it was, the high level of rates on customers'
loans proves little more than the extreme imperfection of the money
market in our sense.2
This imperfection and the abundance of other sources of funds also
disrupts the close relation which according to our model should subsist
between customers' rates and profit ratios in Professor Crum's sense.
1 That question will be taken up in Sec. F, III.
2 The National Bureau average of customers' rates fell at a rate of about 0.8 per cent
per year from 1922 to 1929, while bond yields fell at a rate of 1.4 per cent. See F. C. Mills,
op. cit.t p. 455,
1919-1929
815
Inspecting the curves traced out by the latter (Chart XLV), we find but
unsatisfactory covariation with those rates. We shall understand that
the trough in profit ratios that occurred in 1921 was accompanied by
high loan rates; we shall not be surprised to observe the disruptive influ-
ence of the speculative mania at the end of the period; and we may also
ALL SIX DIVISIONS
1919 I9ZO 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933
CHART XLV. — Profit ratios of United States corporations (see Appendix, p. 1069).
allow for lags. This, no doubt, improves matters, but the relation
remains disappointingly weak.1
2. From the, comparatively speaking, normal course of interest rates
in the United States the German case indeed differs, but not so much as
one might think. Descent during 1924 from the impressive heights at
the beginning of that year — day-to-day money, for example, was at
87.64 per cent for January — has, of course, nothing to do with our process
and merely reflects reaction to the preceding death struggle of the specu-
lation that was an element of the process of wild inflation. Similarly,
the rise in 1929 only reflects the slipping away of foreign balances owing
to the New York stock exchange boom and to a wave of foreign distrust.
But if we connect (see Chart XL) the figures for the first quarter of
1 The covariation of customers' rates with the ratio of net income to capital is more
satisfactory, however.
816 BUSINESS CYCLES
1925 with the figures of the first quarter of 1929 or, still better, those for
the last quarter of 1925 with those of the last quarter of 1930, we get a
fall which may reasonably be attributed to our tendency. The fluctua-
tions within it, i.e., mainly the rise during 1927 and the reaction to it,
also bear interpretation by the cyclical process. While it cannot be
urged against this interpretation that the fall in 1926 was simply a
lagged effect of previous inflows of foreign funds and the rise in 19£7
a lagged effect of the temporary cessation of those inflows in 1926 —
because those inflows were not independent of the German business
situations — it seems permissible to point to the importance of this
ipethod of financing domestic business in explanation of certain obvious
irregularities in timing. Of course, the general level around which rates
fluctuated, though displaying a downward inclination, was abnormally
high throughout: 6 per cent gold mortgage bonds (roughly though not
perfectly comparable to American triple A bonds), which were around
50 per cent of par in 1924, were still but 98.95 in February 1927, when
they began to fall again, under the pressure of foreign sales and of the
strain of the consumers' prosperity. But this is easily accounted for by
the wastages and dissavings of the war and the period of inflation and by
the fiscal policy pursued.
3. In the English curve (Chart XLI; for call rate and the rate on
loans at short notice to the stock exchange see Chart LIV) Kitchins show
well, and there is no reason for not attributing the rise in 1925 partly to
the Juglar phase then prevailing. But a trace of the Kondratieff tendency
is, at best, only visible from the end of 1925 to the end of 1928. The
peak of 1920 and the trough of 1922 are again easily understandable on
short-run considerations — cyclical ones among them — but neither of
them can be taken as a starting point. On the other end, the sharp
rise in 1929 is also easy to understand considering the sensitiveness of a
fundamentally untenable currency situation. We should add, however,
that as shown by the behavior of the rate on stock exchange loans, domes-
tic speculation was also an important factor. But even if we cut off
the postwar hump and thereupon connect the figure for the first quarter
of 1922 with the figure for the first quarter of 1928, we get, in contrast
to the American case, a rising instead of a falling "trend," and even the
trough of 1930 is still above that of 1922. This is of course due to the
policy adopted by the Bank in protecting the pound after it had climbed
up to par, and pro tanto reveals the presence of a disturbance of the
normal course of events as important as, though, of course, not more so
than, anything that affects money rates can be. Not until depreciation
removed the shackles was interest free to move according to its own law.
The high general level of interest rates is, however, not wholly due
to that but also to causes similar to those which were operative in the
1919-1929 817
German case. The index of yields on four fixed interest securities1 fell
from its peak (November 1920: 163; 1913 = 100) to a trough in June
1923. Even at this trough (117) it was perhaps something like 30 points
above what it could possibly have been without the war and postwar
policies. Variations as distinguished from general level are nevertheless
not so far from expectation also in this case. We should have expected
a rise after that trough, owing to the influence of Juglar phases. We
find it, and although it set in earlier and went farther than we might have
expected — the peak which occurred in September 1926 was 139 — and
although other factors were probably more important in bringing it
about, there is again no reason for not attributing it in part to the Juglar
recovery and especially to the Juglar prosperity. The slow fall to
January 1929 (122) also is what we should expect in a Juglar recession.
The differences in the behavior of interest rates in our three countries
are not more obvious than the parallelisms between them. They were
enforced through the mechanism of short balances, which for a time
functioned not less but more promptly than before the war. This can
be seen from the fact that, respectively, from April 1925 and August
1926 until 1930, the Berlin prices of sterling and dollars display very
close inverse associations with the differences between the Berlin and
the London and New York market rates.2 These facts acquire for us
additional significance in the light of the other fact that English, German,
and American cyclical phases were substantially, though not perfectly,
in step.3
II. a. For an idea about the behavior of American system expendi-
ture during the postwar period, we rely on the Federal Reserve Board's
series of debits to individual accounts in 141 cities outside of New York
City. It includes several items which should for our purpose be excluded
and is otherwise not quite what we want. In particular, it is neither free
from the influence of speculative transactions nor strictly comparable
with the series with which we are, nevertheless, about to relate it.4
But still we may put our trust in it with a lighter heart than in the clearing
figures previously used (see Chart XL VI).
The first thing that strikes us is the weakness of the "normalizing
reaction" to the war disturbance. In fact, there is hardly any. The
annual figure for 1921 is about 25 per cent below the obviously abnormal
1 See International Abstract of Economic Statistics, Conference of Economic Services
(London, 1934) or Special Memorandum 83 of the London and Cambridge Economic
Service.
2 See H. Neisser, op. cit.t (Welturirtschaftliches Archiv, 1920). Before August 1926
that was not so in the case of the dollar, because the Reichsbank rigidly held it at 4.20.
8 Bank rates and other points will come up for discussion later on; see below, sub III.
4 This is obvious in the case of the national income, but it should be emphasized that it
is also true of the deposit series.
818
BUSINESS CYCLES
one for 1920, but this fall, unlike that in price level or interest rates, did
little more than eliminate the effect of the postwar boom and the 1919
figure was surpassed already in 1923. No other series shows with equal
force the absence of any significant postwar "deflation" or the fact
that the credit apparatus was by the war expanded for good, which
NET DEM A NODE POSITS
I I I I
ANNUAL AGGREGATE REALIZED INCOME
HOURLY EARNINGS IN MANUFACTURING
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART XLVI.— United States (see Appendix, p. 1069).
stands out particularly impressively if we compare with it the behavior
of the English series of country plus provincial clearings (see Chart LI).
If, then, we accept this as a datum, the rest, viz., an ascent from a monthly
average of 16.6 billions in 1922 to a monthly average of 27.7 billions in
1929, substantially steeper than that of outside bank clearings from 1900
to 1913, also testifies to the unexhausted powers of that apparatus,
though a more modest rate of increase would have been expected by us.
A straight-line descriptive trend would tolerably fit the logs of the figures.
The Kitchins show well and the rise of the fourth Juglar (1925) is indi-
cated by what but for 1928 and 19291 would look like a new level.
1 The figures of these years are of course the ones most influenced by speculative
transactions.
1919-1929 819
Our chart uses the Copeland-Crum figures for realized minus imputed
national income (see Appendix), but no difference would be made in our
rough conclusions if those of Professor Kuznets (op. cit., p. 8, Col. 3 of
Table I) had been used. This series also testifies to the permanence of
the monetary revolution that had occurred but it presents a reduced and,
since this is primarily due to the exclusion from it of capital gains, a truer
picture of it.1 The postwar boom which affected the figures for 1919
and 1920 — the latter, 69 billions, was about 12 per cent above the former
— left national income for 1921 at exactly the level of 1918 (57.4).
Then follows an unbroken ascent, from 61.2 billions in 1922 to 87.6 in
1929. The short-run behavior corresponds — account being taken of
the statistical differences between the two series — to that of debits and
would do so still more if undivided profits were included.
It might seem that the pay-roll index charted displays a still more
moderate increase and thus raises an additional question. But on closer
inspection it will be seen that this is not so. Both national income and
(factory) pay rolls fell 1920-1921 more strongly than they rose in 1919-
1920 (unlike debits). The percentage rise and fall were both more
pronounced in the latter than in the former case. But in 1922 pay rolls
also recovered at a much greater rate than did national income, so much
so that not merely the dip in 1924, which occurs only in the pay-rolls
index, but also its behavior for the rest of the time explains itself naturally
in the light of a reaction to that increase : surveying as a whole the stretch
between (for pay rolls: middle of) 1921 and (for pay rolls: third quarter
of) 1929, we find substantial parallelism. Nevertheless, there possibly
was, apart from the influence of that difference in time shape and from
year-to-year variations,2 also a significant change in the longer run rela-
tion between pay rolls and income which reveals itself in a comparison
made by Professor Copeland of pay rolls in banking and in nonfarm
1 In the long run, or by its "underlying tendency" or its trend values, national income
in terms of current units of currency may be said to measure that component of prices and
values which is, at least immediately, attributable to monetary causes, see Chaps. VIII
and X.
2 Professor M. A. Copeland, National Wealth and Income, Journal of the American
Statistical Association for June 1935, pp. 385-386, notices that the ratio of pay roll to
realized income and the proportion of total income received by the poorest 90 per cent
of the nonfarm population (but gross farm income moved much as income of industrial
workers, see M. Ezekiel, op. tit., chart on p. 140) show similar increments and adds that
from 1918 to 1929 the latter rises and falls with the ups and downs of business. This is so,
but it must be entirely due, first, to the inclusion of the abnormal year 1919, second, to
the abnormally strong increase in wages in 1923, and, third, to the exclusion of undivided
or unwithdrawn profits. Hence, the writer cannot follow Professor Copeland in consider-
ing that finding relevant to "the Hobsonian view [now unfortunately sponsored by, among
lesser lights, Dr. Ezekiel, J. A. S.] that increased concentration in periods of prosperity is
responsible for a disproportionate volume of saved income," however much he rejoices on
820 BUSINESS CYCLES
industry with the comparable part of total realized income.1 This rela-
tion was, at about 73 per cent, in the long run fairly stable before the
war. We should expect and actually find that war expenditure dis-
turbed, i.e., at first, decreased it — we remember that adventitious, in
contrast to entrepreneurial, demand always tends to do that. But reac-
tion against this is observable in 1917, and the improvement in the
bargaining position of labor, the labor policies of the various war boards
and the conditions of the postwar boom carried it much beyond that
prewar level to, if we may trust the figures, about 84 per eent. Some
decline, though owing to the change in Kondratieff phase not necessarily
to prewar levels, was due within the process of general normalization
and may in fact be seen in those figures, although the ratio remained
much above prewar normal in 1929 (77.9). To this and cognate subjects
we shall return below.
b. Corporate accumulations are, as stated above, not included in the
income figures discussed.2 We may use their variations as indicators
of the variations in all accumulations.3 In doing so we must of course
bear in mind that, as presented in the official Statistics of Income,
general grounds in Professor Copeland's refusal to support the implications of that unsound
doctrine. But the theorem that during recession aggregate profits converge toward zero
and during depression tend to become negative is not so very far removed from reality.
1 Professor Copeland's series and the argument that follows in the text involve the
use of two different and imperfectly comparable series, Professor King's from 1909 to 1920
and Professor Copeland's own from 1920 to 1929. If we went on, a third, Professor Kuz-
nets', would have to be used.
a That is to say, they have not been added. But as far as they are not used for the
purpose of increasing cash items (and as far as they really exist), they of course eventually
reappear in individual incomes, although some authors argue as if they were stored up.
3 No estimate will be offered of household savings but only certain indications, because
we simply do not know enough about them, particularly as regards the roughly 89,000
(1929) individual net incomes above $50,000. The most ambitious attempt that has been
made (Levin, Moulton, and Warburton, America's Capacity to Consume, 1934, followed
by two interpretative volumes by H. G. Moulton, Formation of Capital, 1935, and Income
and Economic Progress, 1935) has, statistically and theoretically, been so severely criticized
that it is unnecessary to explain why we do not avail ourselves of the results. In particular,
the author has little, if anything, to add to Mr. H. H. Villard's article on Dr. Moulton's
Estimates of Saving and Investment, American Economic Review for September 1937, which
completely disposes of the contention that there was any excess of net monetary savings
over net productive investment, and thus of one type of oversaving theories as applied to
the postwar period. Concerning another type, it is relevant to note that, as one of the
co-authors of the first of the volumes mentioned, Mr. C. Warburton, has pointed out
(Capacity to Consume, p. Ill, and again in Trend of Savings, 1900-1929, Journal of
Political Economy, 1935, p. 84 et seq.), there was no significant "trend" in the percentage of
savings for the whole period 1917 to 1929, even in the data of the Brookings investigation,
if it were admissible to speak of trends in the case of a series extending over 13 years. There
would be a downward one, or, for 1922 to 1929, a very slightly rising one, if, as we must,
1919-1929 821
they — net incomes minus cash dividends paid to the individuals — are in
some cases the product of arbitrary decisions and in others of irrational
bookkeeping routine. Obsolescence in particular can hardly ever be
adequately taken account of either by the executives with whom those
decisions rest or by the observer. Moreover, the undistributed part of
net income not only fills the function of accumulation in our sense but
also that of an equalization fund, so that they really ought to be referred
not to a year but to, say, a Juglar. Finally, the figures must be cor-
rected for the substantial difference that exists between the usual depre-
ciation at cost and the appropriate depreciation at current prices. This
has been done by Mr. S. Fabricant, whose corrected series for all corpora-
tions except tax-exempt and life insurance companies we are going to use.1
The 1919 boom carried them to a peak that was never reached again —
$3,310 millions, a figure which illustrates the equalization function of
undivided surplus. Whether or not corporate business intended to pre-
pare itself to meet the difficulties that were to come by accumulating
that "reserve," the latter was actually almost wiped out in 1920 (—-$10
millions) and 1921 ( — $3,240 millions). Starting with a practically clean
slate in 1922, we have, for that and the two following years, which about
completed the Juglar recovery, a total of $4,090 millions. The next
3 years, roughly covering the prosperity phase of the fourth Juglar,
added $4,760 millions, over half of them in 1925. This accords well with
expectation. But the contribution of 1928, $2,040 millions and, though
less so, that of 1929, $970 millions, are above expectation2 and must be
included in our growing list of abnormalities which showed during those
we exclude capital gains. (See Chart II on p. 101, inference from which involves, however,
the hypothesis that the percentage of national income saved is roughly equal to the per-
centage of income derived from property.) "With capital gains excluded, this percentage
[of savings] was lower throughout the decade 1920-1929 than during the prewar years
1909-1914" (p. 100). This is in itself sufficient to destroy the case for either an over-
saving or an overinvestment inference from this material. We may take the opportunity
to refer to Mr. W. H. Lough's critique of the Brookings estimates of household savings in
his most helpful book on High-level Consumption, 1935 (Appendix G) and to add that his
own estimates (see below), although owing to the use of a different concept of saving and
for other reasons, far above what savings — which, be it recalled, exclude among other
things sums assembled for residential construction for owner occupancy — in our sense can
have amounted to, yet fail to display an upward trend for 1919 to 1929 (though they of
course increase from 1922 to 1928) and that their percentage share in his grand totals of
realized income displays a falling one.
1 Measures of Capital Consumption, 1919 to 1933, National Bureau of Economic
Research, Bulletin 60, for June 30, 1936, p. 12. The limitations of the material and the
difficulties of handling it are fully discussed in that study, to which the reader is hereby
referred.
2 They are, however, in accord with the Kitchin phases, which in this series show very
well throughout. The dips of 1924 and 1927 are strongly marked.
822 BUSINESS CYCLES
two years. This makes, for the eight years, $1,485 millions per year or,
for the eleven years, $1,084 millions.
There is no point in stressing either figure, because neither 1922 to
1929 nor 1919 to 1929 forms any unit that has any cyclical meaning.
But there is point in stressing their comparative smallness: the first
must, moreover, be judged in the light of the fact that the span to which
it applies does not contain the Juglar depression, which in this case, as
we shall see, more than wiped out the total. Although we cannot in so
short a series speak of a trend in our sense — a result trend — but only of
Juglar and Kitchin fluctuations, it is worth while also to note that a
formal trend through 1922 to 1928 would not display a significant inclina-
tion.1 And, if, from the standpoint of oversaving theories, the figure of
1928 were fastened upon in connection with an explanation of subsequent
vicissitudes, we should ask in reply whether a figure only about half a
billion above the average could, even from that standpoint, be looked
upon as adequate, and why the still higher figure of 1925 did not prevent
a perfectly normal Juglar prosperity and a supernormally active Juglar
recession. The order of magnitude of the total precludes any retreat
on cumulative effects. It was not even enough to prevent a substantial
increase in long-term debt, in spite of the booming stock markets.
Comparing the course of corporate accumulations which has been
just described with the evidence embodied in Chart XLIII, we see what
has often been observed, viz., that there is fair correspondence — though
opinions may differ about the nature and average amount of the lag —
between the variations in those accumulations and in the production
of durable goods, as there also is, let us add, between the former and
expenditure on plant and equipment. This is of course as it should be,
but the inference that might be drawn is much weakened by the fact that
covariation with total production of manufacturing industry is just about
equally good and covariation with consumers' goods' production almost
so. The three — and many other — quantities simply move together
within one process, as, to use Professor Leontief 's happy phrase, soldiers
do in a marching battalion, and there is little justification for picking out
the relation between any two of them and still less for interpreting it
causally.2
1 Cash dividends also display the Juglar and Kitchin phases, but only slightly : they
increase from the last quarter of 1921 to the end of 1929 in an almost straight-line manner.
They were greater by about 3.1 billions in 1929 than in 1922, thus absorbing most of the
net increase in the net revenue, exclusive of taxes, over the period. Corporate accumula-
tion failed to keep step with corporate distribution. The implications of this must, no doubt,
be qualified by taking account of the household savings of shareholders. But what we know
about them does not invalidate the obvious conclusion, viz., that, so far as this source of
income goes, the percentage "saved" decreased as the income itself increased.
2 But we shall not be surprised to find that the more usual "causal" relation has now
1919-1929 823
c. It is of some interest finally to try to form an idea, although it
can only be a very imperfect one, about the relation between total realized
income (which excludes corporate accumulations) and consumers'
expenditure, which will also shed some light on the order of magnitude
of, and on the variations in, net savings of all households combined.
An indication is afforded by the fact that the percentage change in
department-store sales was from about the middle of 1923 to the first
months of 1929 practically equal to the percentage change in nonagri-
cultural income.1 Department-store sales rose more and fell less than
nonagricultural income in 1920 and 1921, but then the indices of both
(1928-1929 = 100) got and kept together until the second quarter of
1929 — after which they fell so closely in step that the curves in 1932
practically coincide — when nonagricultural income rose slightly more
than sales. There is no systematic tendency for variations in sales to
lag behind or to become relatively smaller as aggregate income increases.
The suggestion implied is substantially confirmed, if due account be
taken of consumers' outlay on those durable goods for which department-
store sales are not typical — the main instances being outlay on passenger
cars, which is, and acquisition of homes, which is not, as a rule included
in the estimates of consumers' expenditure. This is done by the figures
compiled by Lough, King, Kuznets, and Warburton.2 Moreover, those
figures indicate, although the fact is veiled by differences in definitions
and classifications, not only that the absolute values of aggregate realized
income and aggregate expenditure on consumers' goods displayed no
been turned around by some theorists of consumers' credit so as to read that stimulation
of consumption will stimulate investment and even " saving" and is, in fact, the life-giver
of the economic process. There is as much truth as there is error in any such statements,
the very inadequacies of which then produce "discoveries" that are pregnant with
policies.
1 This finding is due to Dr. Louis Bean. See Nonagricultural Income as a Measure of
Domestic Demand, by Bean, Bollinger, and Wells, U. S. Department of Agriculture,
Agricultural Adjustment Administration, Agricultural Industrial Relations Section,
June 1937, p. 8. The index of nonagricultural income is not strictly comparable with
our realized income but the contours are the same.
2 W. II. Lough, op. cit., p. 26, chart on p. 27 (Total Outgo, roughly equal to his realized
income, which includes imputed rentals, to be compared with "commodities" plus "intangi-
bles"); W. J. King, National Income and Its Purchasing Power, 1930; S. Kuznets, National
Income (1938), Table 15, p. 53: consumers' outlay (in his sense) fluctuates 1921 to 1929
around 88 per cent, the last year displaying the larg st percentage (90.7), and after rear-
rangement to fit our concepts around 95 per cent of his national income. C. Warburton,
How the National Income was Spent, 1919-1929 (Journal of the American Statistical
Association, March 1935, Papers and Proceedings of the 96th meeting, p. 177, incomparable
with the other estimates as they are between each other) obtains for consumption in per
cent of his national product a figure varying between 79.8 (1919) and 102.6 (1921). Since
1925 shows 85, this is the only series in which the percentage is markedly higher in depres-
sion and recession than in prosperity.
824 BUSINESS CYCLES
tendency to drift apart, as some theories postulate they should, but also
that they did not differ so very much from each other.1
The reader will be inclined to attribute these results to statistical
miscarriages and wonder how either of them can be true, considering,
on the one hand, that we know from common experience that households
do save2 — and, so we are taught, out of increasing incomes more than
out of decreasing ones — and on the other hand, that rearrangements in
the timing of consumers' expenditure — such as is involved in a% shift of
1 Mr. Lough (op. cit., see in particular the basic tabulation in Appendix A and p. 806)
arrives, it is true, for 1919 to 1929 at a total of " saving from realized income" which is
not far short of 100 billions. Relatively minor points apart, this is, of course, due to his
treatment of increases in cash holdings, of all life insurance premiums and of payments
for the acquisition of real estate, and if the total is scaled down to fit our concepts, the
difference between income and expenditure reduces to a quantity of the order of magnitude
of about 5 per cent of the former. We may illustrate the point also by a rearrangement of
Mr. Warburton's figures, op. cit., p. 178. If we deduct from his totals for consumers'
goods and services the item of imputed rentals and if we add to them the item of residential
construction (not all of it ought to be added, of course, but, on the other hand, we neglect
other items which should) we find that the result comes fairly close to the Copeland-Crum
revised series of realized income. The differences (income — consumers outlay) then are in
billions: for 1919, 1.8; for 1921, minus 3.4; for 1923, minus 1.8; for 1925, minus 4; for 1927,
0.8; for 1929, 3.2. This may read absurd (especially the minus figures do) but is not in
the least. That realized income should fall short of consumers' expenditure is not at all
unlikely in a period in which everyone rushed into debt and, incidentally, impugned over-
saving. And there is (though plenty of roughness) no statistical mistake in the comparison
of two substantially independent and comparable aggregates, nor does it follow ipso facto
(this would mean begging the question) that the Copeland-Crum estimates are too low.
No inference can, of course, be drawn from the detail of the behavior of those differences.
They do show, however, that in the present state of our information presence or absence
of net savings is a matter of margins of error and this illustrates the point we wish to make
quite sufficiently. Into the shifts, so interesting from many standpoints, which we observe
between different classes of consumers' expenditure, we cannot and need not go. The
one that is relevant to our purpose, the shift toward durables, has been mentioned and is
in any case obvious.
2 But since earmarkings for the rainy day and for the acquisition of durable goods,
homes included, and nonspending of capital gains all do not come within our concept of
savings — the writer trusts that the reader remembers the reasons we have for this — the
fact is, for the period under discussion, not quite so palpable as it seems. The first two
items will cover much of what is usually considered as the savings of the lower and middle
income groups; the third will cover much of the so-called savings of the middle and higher
groups. We are speaking of this country and a period of an all pervading speculative
attitude and of uncritical optimism. The successful lawyer, doctor, and business execu-
tive, all speculated on the stock exchange and, as long as things went well, were, say, up
to an income of $50,000 exclusive of such gains, under great temptation to look upon their
current earnings as a fund for current expenditure. This may have been different in the
case of seasoned property incomes of the larger and largest sizes. In 1929 there were 513
individual income tax returns stating incomes of a million and over, but only 75 in 1924.
If we look upon the latter incomes as the "seasoned" ones — assuming that recipients of
war profits had also been sobered by that time — we may guess that the prevalent idea
1919-1929 825
consumers' demand toward more durable goods — underspending in deep
depression, increasing replacement of saving by insurance, and direct
taxes paid out of incomes1 must all of them create, if not savings or not
savings in our sense, at least discrepancies between our aggregates, and
raise the figures of realized income — increasingly even — above those of
consumers' expenditure. The answer is simple, however. All those
items were to a large extent compensated by dissaving (mainly the
spending of capital gains) and borrowing, the latter including some that
professed to be for business purposes. There obviously were in all
strata of society very many people who "lived above their means/'2
as is shown by the ready response to any boom or slump, particularly
in stock markets, of the sales of industries catering for "luxury" demand.3
We thus return with added emphasis to the opinion already submitted
for the prewar time in an earlier chapter, viz., that the amount of net
households' savings, as distinguished from firms' accumulation, is being
greatly exaggerated even by the more sober ones of the current estimates.
In support, we may point again to the evidence there is — as, to a lesser
extent, there was for the prewar time — for households' straining their
resources in order to expand their consumption.
about the size of rich men's genuine savings applies fairly well to these, although some
households known to be in that category spent pretty freely on their own consumption,
— the higher the income, the less its relative purchasing power as compared with Europe —
and others on other people's. But those ideas hardly apply to the rest. The drift of the
argument would, of course, not be affected if we had chosen a lower than the million limit.
1 Of underspending there is some evidence for 1921. But otherwise there was np year
of deep depression and, hence, there cannot have been significant underspending. Rear-
rangements of the time shape of expenditure were mostly effected by borrowing, but they
must have had some influence in causing temporary withholdings. The earmarking of
sums for payments of life insurance premiums is, of course, saving in some, but not in other,
cases. State and Federal taxes are (except for 1919) more than balanced by the "public
consumers' goods," public buildings, highways, and streets.
2 That turn of phrase applies strictly to part only of the cases we refer to and becomes
misleading beyond it. Nobody, for example, thought he was living above means, still
less that he was dissaving, if he bought himself a home from realized appreciation of stocks.
3 This is true for the "luxuries" of all income brackets, of course. It is, however, not
less but more true for those of the higher ones. It was not the third-class but the first-
class hotels, for example, which were overcrowded in 1929 and in which a visitor had a whole
floor to himself in 1931 to 1938. Even obviously temporary gains are readily spent and
play a large role in the financing of the consumption precisely of the higher income levels.
That the lower and lower middle classes spent a larger percentage of their income on
consumption than before the war may perhaps be inferred from the expansion in the
production of their luxury goods and is in itself plausible, owing to the growth of social
insurance and to an unmistakable change in attitude. But the former may also be due
to a shift in taste and such indications as we have are hardly trustworthy. See, for exam-
ple, Mr. Warburton's argument in favor of this view, Trend of Saving, p. 97. Nothing
can be concluded from figures such as are there presented without committing a contempt
of court in matters of statistical principle.
826 BUSINESS CYCLES
According to Mr. Lough, short-term consumers' debts increased by
about 4 billions from 1919 to 1929.1 This includes loans against life
insurance policies and federal loans to veterans as well as installment
paper, open accounts, overdue items, loans by loan agencies, pawn-
brokers, personal loans by commercial banks, but not loans against
building and loan shares or home mortgage loans. Very much higher
estimates have been published,2 but the one quoted suffices to indicate
a tendency which becomes still more significant by virtue of the fact that
it is the years of rapidly rising incomes which display the largest increases :
1923 (776 millions: maximum increase), 1928 (609), and 1929 (691).
The most obvious single symptom is the growth of buying on install-
ments, which spread from motorcars and the new household goods —
refrigerators and so on — to clothing and, while it was spreading, involved
anticipation of future income. Exact figures exist only for individual
industries. For the total we have estimates which vary widely : the first
that was based on a careful investigation — that of M. V. Ay res — was
5.7 billions for 1925. It has been criticized, and reduced to 4.875 by
Professor Seligman, still more so by later writers.3 We know, however,4
that in 1927, 23,779 retail stores whose total sales were over 4.7 billions
and which included a number of "cash and carry" establishments, sold
9.2 per cent of the total on installments, 32.2 per cent on open credit
arrangements, and only 58.6 per cent for cash. Motorcar dealers sold
nearly 50 per cent, or according to another estimate 585 per cent on install-
ment; furniture dealers, 57.7; and in the lumber and building materials
trade open credit was given for 90 per cent of the sales.
The role of both borrowings and dissavings in the processes of the
twenties and their relation to the subsequent breakdown is obvious.
But considering the prevalence of the view that consumers displayed
want of alacrity in responding to increase in monetary incomes or, for
that matter, in business revenue and that this was a source of troubles,
1 See op. cit., p. 312. The increase is from 5.4 billions to 9.4, part of it being, of course,
due to carrying charges.
2 The highest that has come to the notice of the writer is that of F. W. Ryan (Internal
Debts of the United States) which for 1929 is 22 billions.
8 These figures refer to total volume of installment sales, not to installment paper out-
standing. Every series of estimates the writer has seen displays a rising "trend" from
1919 to 1929, Mr. Lough's to about double. The significance of this must not be exag-
gerated. A new method of paying for durable goods was simply gaining ground. But
while it expanded, it meant buying beyond the funds available from consumers' income
streams, and this is enough for us. No need to say that bulges occur in the relatively
prosperous years in which people should, according to prevailing theories, be busily saving.
4 National Retail Credit Survey, Part III, 1930, Department of Commerce.
6 C. C. Hanch, Composite Experience of Automobile Finance Companies, 1927. It
had, according to the same source, been 75.5 in 1925. The rapid growth of consumers'
finance companies is in itself a significant fact.
1919-1929 827
the facts glanced at, fragmentary as they are, have an importance in
themselves, irrespective of all theory. Though we do not hold, of course,
that any dire consequences would have followed from a higher amount
or from an increasing percentage rate of savings, we may still note that
oversaving theories would not apply, even if they were logically unexcep-
tionable.
d. In Germany, the estimate of national income which holds the field
is that of the federal statistical office (Reichsamt) . l Since the first
figure is for 1925 (59 billion marks), we have five in all. There was strong
increase to 1928 (73.4)— and a small one to 1929 (73.6)— which would be
somewhat accentuated by the addition of reparation payments which
have been excluded. The national wage bill (which includes all salaries
and is hence affected by the increase, relative as well as absolute, of
salaried employees of all ranks) also increased steadily from 33.7 billions
to 42.6 in 1928 (and 43 in 1929) hence by about 2 per cent more than
national income or (if that increasing weight of salaries be taken into
account) roughly as the latter.2 This too is perfectly normal, although
the movements from year to year reveal various abnormalities, the
most important of which is that the increase in wage bill from 1927 to
1928 was, in absolute amount, almost equal to the increase in national
income. In spite of this, however, that wage bill was even for 1928 not
more than 58 per cent of national income,3 which considering the com-
prehensiveness of the former is anything but high.
In comparing the national income with consumers' expenditure we
strike the same phenomenon, only still more obviously, which we observed
in the American case. That figure of national income includes corporate
accumulation. A careful estimate of consumers* outlay,4 which excludes
1 Monograph (Einzelschrift), No. 14. Estimates of national income at one time
threatened to become a political affair, and it was possible to infer from the figures com-
pilers presented whether they were siding with trade unions or opposed to their policies.
The one we use seems to be entitled to confidence. Its main bases are income-tax and
wage-tax statistics. Double counting is carefully guarded against. This does not mean,
however, that it is strictly comparable to the American series though variations presumably
are.
2 The figures for wages are not quite safe. The wage tax with its many exemptions
and rebates proves on investigation a less valuable ally than one would think. Comparison
with the prewar time when it did not exist is too hazardous a venture to undertake, since
the figure of wage income for 1918 is highly conjectural. This also casts doubts on the
precise significance of the sharp increase from 1924 to 1925 (about 40 per cent in the
industrial pay roll).
3 It is, however, important to bear in mind that the income of wage and salary receivers
was considerably increased by "social rents" and pensions which are not included in the
wage bill and in 1928 amounted to 11.2 per cent of the national income, a largely though
not wholly additive item.
4 Walther Lederer, Was verbrauchen wir? in Die Arbeit, 1932,
8£8 BUSINESS CYCLES
residential construction for owner occupancy,1 nevertheless comes close
to it. The differences (income minus outlay, in billions of marks) for
1925 to 1928 are: 1.3, 1.4, 1.9, 1.7. If taxes paid out of incomes are
taken into account there is a negative figure for every year. This
accords with an investigation of the Institut fiir Konjunkturforschung.2
Now, of course, it will be held — it has been held, as a matter of fact —
that this proves only that the income figures are too low. They may
be. But this cannot be inferred from that fact alone, which it is erroneous
to consider as logically impossible or wildly unlikely. It is perfectly
possible and fits excellently into the general picture of the situation.
People simply borrowed — again in many cases ostensibly for productive
purposes — and dissaved.
For the United Kingdom we use Mr. Colin Clark's income estimates
for 1924 to 1929.3 We start (in billions of pounds) with 3.36, then have
the Juglar jump we expect in 1925 (3,7), an understandable drop in
1926 (3.53), which was roughly made up for in 1927 (3.67) and 1928
(3.64), and end up with some reflex of the 1929 spurt (3.73). The
British income-tax system had (up to 1927-1928, see note below) a
smoothing effect, but considering this, fluctuations show as they should.
There is also a trace of the rising tendency we expect. But the whole is
a very sober affair and characteristically different — so much is clear, in
spite of the lack of comparability of series — from either the American
or the German case. The monetary revolution shows, of course, by
comparison with the last prewar years. But no machine for monetary
expansion had been set up, as' it had in this country, which would, after
liquidation of the war, go on by its own momentum. Pressure on the
aggregate income by the monetary factor is, on the contrary, obvious —
the presence of a component that worked toward continuity with prewar
levels. This would in itself suffice to justify the "prediction" that the
subsequent great depression should be comparatively mild and short.
1 Both series include imputed rentals.
2 Vierteljahrshefte zur Konjunkturforschung, vol. V, No. 4. Mr. Lederer raises various
objections to the result, which is a deficit for each of the 4 years. These objections seem
justified in part. But he only succeeds in turning the deficits into on the average minute
surpluses.
8 See National Income and Outlay, 19S7, p. 88. This, of course, implies that we accept
still another concept. Mr. Clark's most important nostrum, however, the addition to
national income of public revenue from indirect taxation, rates, and so on, we will excuse
ourselves from accepting: the total in Table 35 has been subtracted from the total in Table
37 and so have been undivided profits (Table 85, p. 187). Incomparability with the
American or the German series need not be stressed. Discontinuance in 1927-1928 of
the three-year-average system should be kept in mind. Mr. Clark most commendably
attempts to pierce the veil of the actual tax privileges of agriculture. The British income
tax law does not allow amortization of wasting assets, but does allow the carrying forward
of losses.
1919-1929 829
The wage bill1 includes domestic servants but excludes all, even the
lowest, ranks of salaried labor; hence, it means something altogether
different from both the American and the German series used. It sub-
stantially retained its over-time relation to national income, displaying,
however, a somewhat more pronounced tendency to rise, while the share
of profits was smaller than in 1911, all of which is as it should be according
to Kondratieff expectation. Total wages (in millions of pounds) were
1,399 in 1924; 1,437 in 1925; 1,382 in 1926; 1,492 in 1927; 1,479 in 1928;
and 1,486 in 1929.
Undivided (total) profits,2 which with admirable freedom from preju-
dice Mr. Clark recognizes as "the principal source of savings under
modern conditions," displayed a falling "trend." The highest figure
(186) occurs in 1924; then there was a decline in 1925 (169) and 1926
(134), but imperfectly reversed in 1927 (158) and 1928 (same); 1929 (138)
ushers in the decline, which in England, however, never went into nega-
tive figures (minimum, 1931, was still 28). But again we find that
consumers' outlay comes very close to the income total. Mr. A. E.
Feavearyear3 has made the attempt to estimate average annual national
expenditure for 1924 to 1927. If we add to his list of items of consumers'
expenditure the new houses and the furniture which he includes in saving,
we get about 3.7 billion pounds, which is above any of our income figures
for those years except for 1929. In spite of all differences in conceptual
arrangements, Mr. Clark arrives at a substantially similar conclusion,
though only for 1929 (op. tit., Table 112, p. 252, and Diagram IV and
comments). In fact, "it seems clear that the consumption by the rich
[and "rich" is, according to Mr. Clark, everyone with an income exceed-
ing 250 pounds] was in 1929 about level with their private incomes."
The unavoidable inference is that there cannot have been much, if any,
net saving by private households. Repayments of building loans are
nothing else but payments for a consumers' good bought on the install-
ment plan, the funds of the building societies largely represent not new
1 Op. tit., p. 28.
2 Op. tit., p. 187.
3 Spending the National Income, Economic Journal for March 1931, see table on p. 60.
The way in which Mr. Feavearyear arrives at his figure for savings, which is not less than
400 million pounds, is a good example of the way in which the usual estimates are arrived
at: buying a piano indicates saving; and so does buying a house; average annual amount
of new capital issues accounts for 256 millions — no deduction for duplication or for pay-
ments from ad hoc created funds, etc. — and, of course, there "must" have been expansion of
private business out of income, and there it is. Mr. Feavearyear's result, presented in
another paper (Capital Accumulation and Unemployment, June 1936) — viz., that the
aggregate value of private fortunes in Great Britain, deflated for changes in prices of
assets, has up to 1929 been increasing by about 300 millions a year — though arrived at
after careful investigation, also ceases on analysis to mean what it seems to.
830 BUSINESS CYCLES
savings but shifts in investment. "Security savings" by workmen and
the lower middle classes and net accumulations in life and other insurance
companies there were, of course. But even they — if for the purpose in
hand we agree to treat them as if they were savings — must have been
balanced in part by excess consumption, public and private. We arrive,
therefore, concerning the relation between national income and con-
sumers' outlay, at a result similar to those in the American and German
cases. One significant difference remains, however. Owing partly to
the absence of dazzling hopes for the future and partly to* what the
writer has no other words for than character or moral stamina, there
seems to have been in England no such general rush into debt as there
was in the two other countries.1 And the credit manufacturing appara-
tus proffered much less temptation to it.
e. Profits and wages call for additional comment. Concerning the
former, see for United States corporate profit ratios — percentage ratio
of net income minus income tax to gross income — Chart XLV; for United
States corporate earnings Chart LIV; for German dividends — as rather
doubtful indicators of earnings — Chart XL VIII; and for United Kingdom
"profits"— the most doubtful figures of all— Chart XLIX. We will
confine our comments to this country, because American research on the
subject, much superior to any other, has most nearly succeeded in bring-
ing out the main contours of this complicated tangle of facts.2
In spite of the excellent work done we are, however, as yet far from
seeing clearly, and any inference has to contend with disheartening
margins of error. Not only is the raw material the product of book-
keeping processes which unavoidably deviate, to an unknown extent
that is sure to vary as between industries and individual concerns, from
1 Housing might be considered an exception; but subsidized and safeguarded and run
on the your-rent-will-buy-your-house plan as it was, it may fairly be looked upon as a
special case.
2 The author's primary obligation is to the work of Professor Crum, in particular, to
Corporate Earning Power, 1929, but also to his many papers on the subject. In the second
place, Professor R. C. Epstein's great investigation has been of great help: Industrial
Profits in the United States, National Bureau Publication, 1984, with an introduction by
Professor Mitchell. But on the theory brought to bear on the interpretation of the data cf.
Mrs. Tappan Hollond's review in the Economic Journal, 1935. In the third place, mention
should be made of L. H. Sloan Corporation, Profits, 1929; S. H. Nerlove, A Decade of
Corporate Incomes, 1932; R. T. Bowman, The Statistical Study of Profits, 1934, W. A.
Paton, Corporate Profits as Shown by Audit Reports, National Bureau Publication, 1935;
and Professor F. C. Mill's Economic Tendencies; as well as to a number of papers by L.
Bagwell, L. R. Robinson, and others. The only way toward real insight would, of course,
be the detailed study of the life history of individual concerns, part of the material for
which is to be found in the discussions of annual statements in the financial press. For
Germany the Bilanzanalysen of the Deutsche Volkswirt are an excellent source of this type
of information.
1919-1929 831
the actual state of things; and not only does it largely fail to represent at
all adequately the limbo in which dwell the abortive and short-lived
attempts at enterprise; but even if that were not so, profit figures would
at their best give but a medley of economically heterogeneous elements,
of which entrepreneurial profits1 are only one, though we may perhaps
hope that, being the most active one, they will show up better in the
fluctuations of the aggregate than on other grounds we have a right to
expect.
Since these difficulties particularly interfere with the meaning of the
various ratios that have been computed, we had better start with aggre-
gate net income of all corporations before payment of income taxes and
dividends.2 The violent fluctuations at the beginning are due to war
effects and to the Juglar depression (1919, 9.3 billions; 1921, 0.64), but
from 1923 (6.64) to 1929 (9.13) we have what looks like an almost
steadily rising "trend," substantially similar to that in total national
income: the per cent relation of corporate income to realized national
income (Copeland-Crum series) plus corrected corporate accumulations,
was fairly constant — for 1923 to 1929 it was 9.2, 8, 10.2, 9.5, 8.3, 10,
10.3. Within this contour, cyclical phases are clearly recognizable both
in absolute figures and in percentages. The rise of the fourth Juglar is
particularly well marked: from 1924 (5.74 billions) there is a character-
istic jump to 1925 (nearly 8 billions) and after that some tapering off
(1926, 7.84 billions; 1927, 6.84), interrupted by the abnormalities of
1928 (8.67) and 1929. But this shows what we are to think of that
"trend." On the one hand, 1923 and 1924 belong to a Juglar recovery
that is followed by a Juglar prosperity; on the other hand, the two
abnormally prosperous years happen to be the last ones of the series.3
Of the other ratios — we do not revert to the profit ratio in Professor
Crum's sense — the lowest is the earnings ratio or percentage of net income
1 As we have seen in Chap. Ill, this element tends, in every individual case, to converge
toward zero, but does not otherwise harbor any tendency toward equalization.
2 Exclusive of tax-exempt and life insurance companies. See Ebersole, Burr, and Peter-
son in Review of Economic Statistics for November 1929, and Fabricant, Recent Corporate
Profits in the United States, National Bureau of Economic Research, for Apr. 18, 1934.
Dividends received from other corporations are excluded.
3 It is, hence, inadmissible to use that "trend" as an indication of long-run tendencies
inherent in capitalist evolution and, in particular, as an indication of long-run tendencies
in distributive shares. Even disregarding the necessity of correcting totals for increase in
investment, the true picture emerges only if the subsequent depression be included. But
it is wholly misleading to use the "trend" in total dividend payments — let alone dividends
plus cash value of rights— for either purpose, because, as stated before, they absorbed an
increasing percentage of total corporate net incomes. This is interesting evidence about
the alleged saving propensities of the period, but is otherwise a purely intracapitalistic
affair which has no bearing on the relative fortunes of social classes in the distributive
process.
BUSINESS CYCLES
to total assets, which, since not all corporations file balance sheets, is a
matter that involves much estimating. For the three years for which
Professor Crum calculated it, 1924 to 1926, it was, taking all divisions
together, roughly between 2 and 3 per cent, manufacturing industry
leading with about 5. For 1926 in particular, the aggregate return on
total assets was 1.98 per cent, for manufacturing industry alone, 4.36.1
The latter figure just about equals the yield of highest grade bonds in
the same year — a fact which is of some importance although what we
have before us is an average and not a marginal quantity and although
profit and interest are obviously not independent of each other. Now,
whatever ratio may be relevant for other purposes, this one is relevant for
ours. And since it is quite as low as we should expect it to be in the
downgrade of a Kondratieff, we conclude that there cannot have been
much of a general "profit inflation,"2 for this would have shown precisely
in the general level of income per unit of assets. This may have been
different in 1928 and 1929, but such indications as we have do not sug-
gest that the earnings ratio was appreciably higher. In some cases it was
lower than in 1926. The reason why it was so low is of course that all
losses made by reporting — i.e., still existing — corporations enter into it,
as well as the gains. Taking only corporations reporting gains, Professor
Crum obtains, for 1926, 6.95 per cent in the manufacturing division and
3.66 in all. But according to his showing, almost half of all corporations
worked at a loss or at no profit or at practically none.3 It follows that
expectation from our model is verified, not only as to the size of the
earnings ratio, but also as to the reason for that size. For it is obvious
that such prevalence of losses or, in a business sense, inadequate returns,
1 Professor Crum does not add interest charges: as stated, he considers net income.
We have done the same in the preceding paragraph. In fact, it is open to question whether
to add interest (or only interest on long-term debt) is conducive to a correct impression
of the rate of profits. It might just as well be argued that what we ought to do is to deduct
also interest on owned capital.
2 That term is not used in Mr. Keynes' sense, Treatise on Money I, p. 155, but merely to
indicate all cases in which profits, by virtue of being obviously abnormal, might be taken
as a symptom, consequence, or cause of other abnormalities or disturbances, e.g., as a
cause of overexpansion of output or investment or as a symptom of monopoloid under-
utilization of resources.
3 This finding is substantially confirmed for the whole period by Professor Epstein's
investigation, see op. cit., p. 457: among manufacturing corporations the "with net income"
groups amounts to roughly 60 per cent of the total which was, 1919 to 1928, between
70,000 and 90,000. "Thus in all years from 1919 to 1928 the number of manufacturing
corporations with net incomes runs from about 50,000 to 55,000 a year, except in 1921 when
the figure was slightly less than 40,000." Hence, if Professor Epstein rightly takes excep-
tion to the "common impression • • • that about 50 per cent of the corporations in the
country lose money," we ought to add that this common impression expresses a very
important truth much more nearly correctly than it is usual for common impressions to do.
1919-1929 833
is but another symptom1 of the vigor with which our competing-down
process did its work and of the fact that the prewar process still persisted :
from our standpoint, all this was entirely normal and merely one of the
aspects of economic "progress."2 If it were possible to go into details,
further conformity with expectation would reveal itself.3
Next we will glance at rates of return on book value of stock equity
of "all" corporations. Mr. Fabricant's figures4 for 1927 to 1930 will
suffice. In the grand total they are 5.3, 6.2, 6.2, 2.2; for manufacturing
alone 6.2, 7.6, 8.3, 2.6 per cent. They differ from earnings ratios by
virtue of the agreements entered into with one another by the various
groups of capitalist claimants, and they have, hence, for our purpose,
little importance in themselves. But they may again be compared to
bond yields. Discarding rather revealing details, we again get the
impression that under this aspect even these rates do not suggest "profit
1 There were of course corporations which existed for the very purpose of working at a
loss; others, the profits of which, or more than that, went to the executives; and still others
which, being subsidiaries of some other concern, were on principle selling at cost or
even at a loss. But it will hardly be averred that cases of these types were significant
enough to interfere with our conclusions. Nor can it be objected that the evolutionary
process failed to eliminate the antiquated, ill-conceived, or otherwise unsuccessful elements.
We have in earlier chapters dwelt on the reasons why losing concerns often "hang on"
for some time. But the going-out-of-business rate was considerable throughout, only
it does not show in this kind of material. Finally, it cannot be objected that the process
failed to work in the sphere of bigger and of big corporations which were individually suc-
cessful more or less all the time. For the relation between size and success, although
not one of proportionality, works not only one way. It may be added that, since the
competing-down process takes time (if it did not, economic life would be a continuum of
catastrophes) we shall not share Professor Epstein's astonishment at finding that sub-
stantially the same divisions and concerns were, throughout, at or near the watershed of
loss: all that this proves is that business life is not a game of chance and that profits are
not adequately described as windfalls.
2 What was, from the standpoint of our analysis, a highly "normal" state of things does
not seem so to other economists. Professor Crum seems to consider the facts that "num-
bers of enterprises • • • are dragging along with a very low rate of return on their prop-
erty," and that "a considerable share of the total gross corporate business is done at a loss"
as a reason to doubt "the long-run healthfulness of corporate industry." It is submitted
that attention to the logic and rhythms of the capitalist process completely removes any
such doubt, although that eminent economist was perfectly right if he intended his guarded
statement to imply a prediction of impending vicissitudes. Other economists are in the
habit of taking the facts discussed as proof of the wastefulness of competition or of capital-
ism in general. If, however, they mean more than a triviality, they are wrong. Those
losses are not waste or per se indications of waste in the sense that the social organism as
a whole gets nothing in return, provided they are placed in their proper setting. And
they would, of course, be unavoidable also* in a planned economy.
3 Especially in a study of the concerns that enjoyed more than average returns. Mr.
Sloan's study (op. dt.) of 545 big corporations sheds some light on this and is recom-
mended to the reader.
4 Op. cit., p. 8; net income is taken "after tax."
834 BUSINESS CYCLES
inflation." In 1927 AAA bonds yielded a little over 4.3 per cent, in
1929 nearly 5. European prewar experience — American bookkeeping
methods before 1909 do not permit comparison with American prewar
experience — would, very roughly, support the view that such a margin
was about normal. And the 1930 figure is also "normal" for incipient
depression. Textiles, leather, and rubber display decline from 1927 to
1929; food-beverages-tobacco, paper and pulp, stone-clay-glass are exam-
ples of comparative stability; chemicals and metals, of strong increase
• — surely nothing to be surprised at.
Finally, Professor Epstein's figures for percentage profit (net income
plus interest on funded debt, "before tax") to total capital (stock, com-
mon and preferred, surplus, undivided profits, funded debt, but not other
debt) are for 1924 to 1928 11 5.9, 7.5, 7.2, 6.4, 7.3, but for the sample of
3,144 corporations, 9.2, 10.7, 10.9, 9.4, 10.2.2 The latter may serve to
illustrate what we should have taken as evidence of abnormally high
profits if this had been the result for all corporations or even a random
sample. The 2,046 manufacturing corporations display still higher
figures but, for 1926, the modal percentages are from 5 to 9 and the high
average is as much due to the inclusion of all peak successes as to the
almost complete absence (less than 4 per cent) of cases of loss. Cyclical
fluctuations, the rise of the fourth Juglar in particular, are much in
evidence : it is interesting to notice that, contrary to a prevalent impres-
sion, 1925 (or 1926) was the most profitable year, though, of course,
overshadowed by 1919 and, before that, by 1916-1917. The range of
variation as between the 106 constituent groups and, within them, as
between concerns is as wide as we should expect.3
/. Gathering up the threads of our analysis of wages and employ-
ment (this section, I, a, II, a, and II, d; Sec. C and Sec. E9 passim;
Chap. XI) we will now discuss the behavior of wage rates4 and the effects
1 Op. cit. p. 50.
2 Op. cit., p. 53.
8 The limited purpose of this sketch makes it impossible to do justice to the rich lode
of information to which Professor Epstein's work has opened access. It must, however,
be stressed again that there is no reason why, in a period such as 1919 to 1929, and in a
sample constituted as this is, extra gains should balance extra losses. Nor is it easy to
see why Professor Epstein should so categorically aver that the differences in earning rates
of different industries "cannot be regarded as differentials due to the rent of superior busi-
ness abilities" (p. 582). Nothing is further from the present writer's mind than a wish
to defend that antediluvian turn of phrase, which moreover points in the wrong direction.
But in looking at Professor Epstein's list of particularly successful industries, one may well
wonder whether success has not very much to do with the quality of the products of some
and the dash of the advertising of others. And what is there so very "unlikely" in the
statement that, on the one hand, able men take to promising jobs and, on the other hand,
jobs become promising in the hands of able men?
* We use rates wherever possible because they come nearest to representing the price
1919-1929
835
that may have emanated from them. American facts are presented in
Chart XL VII.
1. In appraising the evidence contained in these curves a number of
limitations must be borne in mind which impose extreme caution in
1919 1920 1921 1922 1923 1924 _ 1925 . 1926 . 1927 1928 _ 1929 1930 193L_ 1932 _ 1933 1934 1935
CHART XL VII. — United States (see Appendix, p. 1070).
drawing inferences. Pay rolls, which have been plotted again and can
now be compared with both " corrected J>1 and "real" pay rolls, and
employment represent, of course, pay rolls and employment in manufac-
turing industry only.2 Hence, even disregarding that they do so imper-
of a definite quantity of labor. Weekly earnings per workman or per employed workman
or per working-class family are, of course, more important for considerations about welfare,
social justice, and so on. They would be more important than rates also for us if we could
enter into the question how the working class fared during what was in this country a
span (though not one which can be taken as either typical or near average) of relatively
uninhibited capitalism. This question must, however, remain outside of our range.
We will merely note that in this country actual weekly earnings in manufactures, which
had risen to a peak of .$29.48 in 1920, suffered a "normalizing" decline to $23.23 in 1922
and then rose again to within 10 per cent of that peak ($27.36) in 1929. See L. Wolman,
National Bureau Bulletin for May 1, 1933, p. 2.
1 The same terminology is used in this argument as in our discussion of prewar wages and
employment.
2 For a closer analysis, see W. A. Berridge, Review of Economic Statistics for November
1930.
836 BUSINESS CYCLES
fectly (as is obvious from the description in the appendix), they do not
indicate the course of total income and total employment of the working
class but only what happened in a sector of the national economy. The
full importance of this becomes evident if we recall that the develop-
ments in manufacturing industry were instrumental in creating additional
employment and labor income in certain other sectors — the sector of
"services" in particular — so that, in comparing manufacturing pay rolls
with say value of output of manufactures or, as we did before, with
total national income, we are not isolating a self-contained relation but
cutting through a nexus which is at the same time more comprehensive
and more relevant.1 Moreover, all the series used are open to objection
on various counts, that of cost of living, in particular, to the old one that
the improvement in quality of the commodities which entered the Ameri-
1 Uncritical comparisons of wage bills or wage rates with other aggregates, often used as
a basis for inferences and value judgments about distributive shares, are a frequent source
of much coarser errors than the one alluded to in the text, if not completely meaningless.
For instance, there is, for the purposes of the study of cyclical fluctuations, some sense in
comparing fluctuations in wage bill with fluctuations in the Census Bureau figures of total
transactions in manufactured products which is sometimes misleadingly referred to as the
gross value of those products and differs from value added plus value of domestic raw
materials plus value of imported producers' goods mainly by the interfirm transactions in
commodities. But there is no sense in instituting this comparison for the purpose of
measuring labor's share. Nor is there any sense in comparing average per capita earnings
of employees with the sum total of, say, stockholders' cash receipts or, still worse, with
stockholders' cash receipts plus stock dividends, or with dividends plus undivided surplus.
Much erroneous argument is due to the confusion, adverted to in Chap. II, of product per
man-hour and the productivity of labor. This accounts for the habit of speaking of varia-
tions in the former as if they were variations in the "social contribution" of labor, and of
comparing them with variations in wage rates on the hypothesis that proportional covaria-
tion between the two is in some sense normal and that deviations from it are something to
be wondered at or to criticize. We are not attacking any ideals that may be the premises,
but the economic errors that are at the basis, of judgments about, and sometimes even of
mere presentations of, the facts — errors which very simple considerations on Boehm-Bawer-
kian lines should be sufficient to dispel. The boldest attempt at introducing econqmic
meaning into discussions of statistics of that kind has been made by Professor P. Douglas.
None of the many reservations that have to be made on many counts, detracts from the
merit of his Theory of Wages, 1984. It should be added, however, that many economists
who do not, in dealing with wage questions, forget their courses in elementary theory,
yet fail to realize the importance of the fact that the fundamental theorem about marginal
value productivity of labor is an equilibrium proposition that would at best apply (approxi-
mately) in neighborhoods of equilibrium, but cannot in the intervals between them.
Profits in our sense precisely arise and vanish in these intervals, hence, do not bear any
definite relation to "productivity wages." Moreover, if the reader will forgive a triviality
which is always obscured in popular discussions, variations in productivity wages and the
corresponding variations in output have much less to do with what in common parlance we
understand by the personal efficiency of the workmen or with what in this sense may be
attributable to their efforts, than with the variations in the amounts of the other factors
applied and with the changes in "methods of production."
1919-1929 887
can working-class budget — food, clothing and what we have termed
gadgets of modern life in particular — was one of the outstanding features
of the period. Finally, wage rates varied so greatly between localities
and, in the same place, between industries and firms that, as we have
seen in an earlier chapter, speaking of one national wage rate and its
variations might well be thought inadmissible, even if the concept of a
national rate per hour were unexceptionable in itself. But it is not.
Piece rates, bonuses, and so on drive a wedge between hourly rates and
hourly earnings, which makes it impossible to infer the one from the
other, still more rates per hour from earnings per week or pay rolls.
In using the terms hourly earnings and hourly rates as if they were syn-
onymous, we are guilty of a serious, though very common, misdemeanor,
in extenuation of which we can only plead that this difficulty was not so
great from 1923 to 1929 as it afterwards became.
Expectation as to wage rates in the Kondratieff phases obtaining
during the period is for a moderate rise in money and a substantial one
in corrected and real rates. If we start 'from the middle of 1923, this
is, on the whole, what we find, and the Juglar and Kitchin variations
are also recognizable.1 In other words, there is, as far as that piece
of evidence goes, no reason to suppose that variations in wage rates
were anything but normal in the sense that they did not interfere with
the expansion, or contribute to the subsequent contraction, of business
volumes by being "too high": the only question seems to be whether
they were not "too low." Again, however, we observe, surveying the
whole period and comparing the level at which rates moved with that
of the last prewar years, not only the traces of the monetary revolution,
as we do in other comparable quantities,2 but also the fact that the down-
ward revision after the postwar peak was comparatively small and only
temporary. Hence, it is at least a possibility that the war and the
postwar boom left labor a relatively dearer factor of production than it
had been before. That it became a relatively more expensive consumers'
good is, of course, beyond reasonable doubt. Suggestion to that effect
is present in nearly all available wage-rate series, but we will content
1 Better than in the series used in our chart (see Appendix) the course of wage rates
shows in the new composite wage index of the Federal Reserve Bank of New York, the
latest improvement of Mr. Snyder's index (see Monthly Review of Credit and Business
Conditions, for Feb. 1, 1938, p. 12, where that comprehensive index, 1926 = 100, has been
charted on a natural scale). The increase in money wage rates between 1923 and 1929
is quite appreciable, the rise of the fourth Juglar shows well. This should be noticed,
because there is a prevalent impression to the effect that from 1923 rates did not rise at all.
2 Price levels, contrary to a widely held opinion, are not among those comparable
quantities. The reader recalls that expectations for commodity prices and wages differ
fundamentally and that there is not the same reason to expect that wages should fall in
recession as there is that prices should.
838 BUSINESS CYCLES
ourselves with one, the basic wage rate for common labor in the Pitts-
burgh district, as reported by the U. S. Steel Corporation.1 This rate
was slightly under 20 cents per hour for 1913 (monthly average) and
nearly 51 for 1920. It declined in 1921 and part of 1922, but was back
again to 50 cents in the monthly average of 1924, after which it remained
constant through 1930. 2 It does not make a great deal of difference
whether we say that the irregularities at the threshold of our period veil
the extent of the rise in money wage rates or, as we put it above, that
they raised the level from which the "regular" developments start.
In both cases, the course of wages must be considered as resulting from
two component tendencies — one which tended to depress and one which
tended to raise them. On the whole, however, the latter way of express-
ing American wage facts seems preferable for our purpose, because it
brings out more clearly that that "level," created by war and postwar
irregularities and, owing to the resistance offered by the environment to
downward revision, substantially left as it was, was a noncyclical and
almost constant force throughout.
2. Even if we look at wage rates in this light, we shall hardly find
any reason to recede from the result provisionally stated above, viz.,
that wages did not hamper prosperities or cause relapses by being "too
high." There is neither any evidence the writer can think of that they
did nor any expectation to that effect from theory — within limits, a
wage level that persists through a decade becomes a datum to which the
system will in general adapt itself without changing its mode of working.3
1 The series has much to recommend it but is not quite consistent in meaning. In
particular, these rates applied to working days of different length, which is not only relevant
to daily earnings, and 1918 to 1921 also reflect the higher payment for overtime. The
series of hourly rates paid for common labor in road building (Bureau of Public Roads)
gives a somewhat different picture. In 1915 the rate was 20 cents. It steadily rose to
49 for 1920, declined in 1921 and 1922 to 82. Then it was at 38 cents for 3 years and rose
in 1928 (maximum) only to 40. Hourly wages in soft-coal mining continued to fall after
1922 (85.3 cents; 1929, 65.9).
2 The index referred to in the last note but two rises spectacularly in 1919 and (three
quarters of) 1920, then falls sharply through 1921 and, though recovering substantially
in 1922 and especially 1923, does not come as near to the all-time peak as the rate discussed
in the text. On the other hand, it continues to rise and ends up in 1929 with a value higher
than that peak. The implication is, hence, fundamentally the same.
3 Qualifications of this proposition will be evident from the rest of the paragraph.
It would not be true, of course, of all deviations from the "normal" course of things. But
it should be noticed that to some extent it agrees with an opinion that is at present held by
many economists, who would, however, neither stress the long-run-level aspect nor accept
the qualifications. Professor Myrdal's and Mr. Keynes' teaching may be referred to in
illustration. To put it differently, the above proposition formulates an element of truth
that is contained in the — otherwise untenable — opinion that the absolute level of monetary
wage rates does not matter. There is another element of truth in it which is, however,
trivial: changes in the absolute level of monetary wage rates do not matter if all other
monetary magnitudes and expressions move uno actu and proportionately.
1919-1929 839
But this merely means that the general complexion of successive business
situations was not substantially affected by it, i.e., not only that pros-
perity remained prosperity and recession remained recession, but also
that all phases were presumably as "intensive" as they would have been
with a somewhat lower level of wages. It does not mean that there were
no effects, in particular on employment. For one of the ways in which
the system would adapt itself to a high level of wages consists precisely
in making of the dear labor factor or labor commodity as economical
a use as possible. We shall, therefore, suspect that during the period
an additional source of unemployment may have been present to swell
the total: we have already noticed that the Kondratieff phases into which
our period falls, would "naturally'* display considerable technological
unemployment; moreover, it goes without saying that, especially in
1921, but also in other years — 1927, for instance — there must have been
cyclical disturbance unemployment of the kind that is not directly
traceable to innovation, as well as unemployment due to random causes,
such as the Mississippi flood and others; we may now have to add unem-
ployment of the type which we have called vicarious.
Now, our factual knowledge and analytic powers1 being what they
1 We are, moreover, at a disadvantage as against such theoretical reasoning on the
subject as has recently been offered, since we cannot state the theoretical case without
expanding this subsection into a treatise on wages. But it should be observed that the
above argument about economizing dear factors is not open to the objection that any
increase in the expenses of production which is due to increase in wages is at least com-
pensated by an equal increase in producers' revenue. For, even if that were always so,
the individual firm would still have a motive to react to an increase in wages by a reduction
of the labor employed per unit of product. The consequent rearrangement of its combina-
tion of factors cannot be neglected for a period of the length of a Juglar, or even, in many
cases, of a Kitchin. And this is as true under conditions of imperfect competition as it is
under conditions of perfect competition. It should be added that in other respects also
prevalence of conditions of imperfect competition affects the argument about the conse-
quences on employment (and output) of variations in monetary and real wage rates less
than might be thought. The elegant argument presented by Dr. P. Sweezy at the 1937
meeting of the American Economic Association (see American Economic Review, Supple-
ment, March 1938, p. 156) may serve as an example. The present writer entirely agrees
with his emphasis on the importance for employment of cyclical shifts in the demand curve
for products (p. 157) and has been at pains to stress this (Chaps. XI and XII) both in the
matter of wages and of interest rates. It is partly for this reason that the short-run
influence on cyclical phases of variations in wage rates has not been assigned a more
important role in the text. But so far as Dr. Sweezy's argument is based on the proposition
that with imperfect competition firms will not react to shifts in their cost curves, because
this involves raising the prices of their product, which would drive business away, or
lowering them, which is self-defeating since it will induce competitors to follow suit, it has,
while logically correct, little claim to being considered more realistic than others. For it is
obviously more likely that competitors of the firm that contracts output and raises prices
because of an increase in wages, will do the same than it is that they will try to conquer the
field from which the latter now retires; and it is no less obvious that in a highly "dynamic"
society the motive for expanding output and, for this purpose, reducing prices is but little
840 BUSINESS CYCLES
are, it is impossible to speak with confidence and to offer anything like
proof. We are inadequately informed about the facts of unemploy-
ment in this country until the unemployment census of 1930.1 From
an estimate that has Professor Wesley C. Mitchell's sanction,2 no "trend"
can be deduced that would have any meaning. This accords with pre-
vious findings (see above, I, a, 3). For 1920, the annual figure already
displaying the influence of the slump, unemployment was 5.1 per cent
of nonagricultural earners of wages and salaries. The slump figures
(1921 and 1922) are 15.3 and 12.1 per cent. The absolute amount corre-
sponding to the former percentage (4.27 millions) probably "surpassed
all previous records," but there is no reason to believe that the percentage
itself did. In any case it fell back to 5.2 per cent for 1923 and rose to
7.7 in 1924. The years 1925 and 1926, with 5.7 and 5.2, indicate an
effect of the Juglar prosperity, but for 1927 we have again 6.3 per cent.
The writer thinks that the figure for 1928 was higher than that and the
figure for 1929 certainly not lower than that for 1928. It is still more
difficult to interpret these figures than it is to trust them. They cer-
tainly include very little of typically spurious or "malingering" unem-
ployment, since there was no government dole. But the nomadic
habits of the American workman arid the high level of earnings that made
it easy to tide over a short spell of unemployment and even to look upon
it as a holiday, suggest that those figures may include a nonnegligible
number of cases that bordered on voluntary unemployment. For these
and other reasons, normal unemployment in this country always has
weakened by the knowledge that competitors will move in the same direction — the stand-
ard instance is once more provided by the industry that typifies entrepreneurial behavior
so well, the motorcar industry, the leading firm in which repeatedly reduced prices, although
it must have known that, in the static sense adopted by Dr. Sweezy at this point, this
measure would be self-defeating. We emphasize this because so much has been made,
in the discussion of these problems, of special cases and because the theory of imperfect
competition has been so fertile in such cases, which, though interesting, only serve to con-
fuse the broad issue.
1 And even the results of that census have been severely questioned by competent
critics.
2 See Recent Economic Changes, 1929, II, p. 879 and the chapter on labor. The
figures are intended to convey an idea of how great unemployment has at least been in
each of the years 1920 to 1927. If it actually was greater, our argument in this and the
next paragraph would apply a fortiori. But it is not certain that it was. While one party
to the discussion, the administration included, may have been resolved to see as little
unemployment as possible, the other party to the discussion was no less resolved to see
as much as possible. The tendency to exaggerate is, in some estimates, as obvious as is
the motive for it: in a period in which it was clear that the majority of the nonagrarian
population was thoroughly contented with the results of the capitalist process, the critic
had little else to fall back upon and naturally made the most of it. A great deal depends
upon definition. But, however we define, there was certainly much "unseen unemploy-
ment" in addition.
1919-1929 841
been higher than in Europe. If, in addition, we take account of our
expectation from the prevailing Kondratieff phases, the conclusion can
only be very tentative that an estimate of minimum unemployment
which, even in years of buoyant activity, never fell to 5 per cent, indi-
cates the presence of still another factor that made for unemployment.
Direct observation of everyday practice, however, not only confirms
this conclusion but also reveals what that factor was. The effort to be
as economical as possible of hired labor, which has been described above
as the immediate consequence of relative dearness of labor, was in fact
an obvious feature of the period and among the first things to strike a
foreign observer of American industrial and private life. As to the
former, it is true of course that labor-saving changes in methods of pro-
duction would have come about in the normal course of our process quite
irrespective of the level of wages. But many of them were conditioned
by, and most of the noninnovating rearrangements of the combinations of
factors of production which occurred in 1920, 1921, and 1922 were clearly
reactions to, a price of labor that was high relatively to that of other factors. l
Money cost of labor per unit of product declined substantially to 1925
and then again from 1927 to 1929. 2 But it is easy to see that, given the
high long-run elasticity of the individual firm's demand for labor, this
fact strengthens the case instead of weakening it. Two circumstances
must, however, be mentioned which worked in the opposite direction.
First, the wage rate under discussion is an average. Actual rates tended,
on the whole though not always,3 to differ from each other in a manner
that would mitigate effects on employment: they were as a rule lower
in weaker geographical or industrial sectors — in the South, for instance,
or in the bituminous coal industry — and higher in others that were able
to bear it. Second, a wage rate high enough to induce substitution of
labor by other factors — or labor of the "manual" kind by other labor —
will at first tend to increase employment, because it engenders additional
demand for labor-saving devices, most of which have themselves to be
produced. During our period that part of current demand for labor-
1 We shall observe the same phenomenon in the processes of recent years. To the
increases in wages that occurred in 1936 and 1937, for instance, industry immediately
reacted by "rationalization," which in many cases almost succeeded in keeping labor cost
per unit of product constant. Southern farmers, not immediately affected or not so much
affected by those increases, are already preparing to meet them by mechanization.
2 See data for 62 industries, F. C. Mills, Economic Tendencies, p. 404. There was an
interruption of the decline, which in those data shows in the figures for 1927, that accords
with our expectation for a Juglar prosperity.
3 Massachusetts, or even the whole of New England, is a conspicuous geographical
exception; and the fate of her industries from 1923 to 1929 shaped accordingly. See D. H.
Davenport and J. J. Croston, Unemployment and Prospects for Reemployment in Massa-
chusetts, Publications of the [Harvard] Graduate School of Business Administration,
Business Research Studies, No. 15, 1936, Chart R on p. 63.
842 BUSINESS CYCLES
saving equipment which constituted a reaction to the prevailing wage
level no doubt helped to keep up total demand for labor. This possible
source of treacherous "verifications" of high-wage theories should always
be kept in mind whenever the effects of wage rates on employment are
being discussed.
That the price of labor considered as a consumers' good was very
high relatively to the prices of other consumers 'goods and that households
reacted accordingly1 is too obvious to detain us. It should merely be
added that this reaction went even further than appears at first sight.
The mechanization of the household was so much more successful in this
country than in others because it not only economized labor but also
made it easier to dispense with hired labor altogether. Even the pros-
perous American family thus learned to perform services for itself which,
but for the rate of wages and especially for the rate of wages per unit of
service, it would have delegated to nomnembers. It also learned to
substitute for enjoyments which involve the direct employment of labor
others which do so less or not at all. Thus the American style of private
life was powerfully influenced, dominated perhaps, by the level of wages.2
Part of the demand for durable consumers' goods which was so important
a feature of the twenties was ultimately nothing else but a flight from
labor arid, in particular, from the relatively most expensive kind, viz.,
manual labor.
It might be asked how the level of wages can have influenced employ-
ment and yet not have — appreciably — influenced what we have called
the general complexion of business situations. The answer is, first,
that the high wage rates themselves largely compensated for the effects
beyond those on employment. From our impression that more labor
would have been employed in American manufacturing industry if wage
rates had been lower than they were, it does not follow that its pay roll
would have been much different from, still less that it would necessarily
have been higher than, what it actually was (see below). Second, part
of the additional employment which would have accompanied lower
wage rates, would not, even if it had brought about a substantial net
1 Households also reacted directly to the price of labor in its productive uses. The
relative increase in the use of consumers' goods requiring less labor than others, and the
relative decrease in the use of consumers' goods requiring more, may be instanced by
the shift of consumers' demand from custom tailored to ready-made clothing.
2 It will be held by some that that style of life is in itself an achievement in efficiency
and by others that it spelled increasing democratization and a moral progress. Certainly.
The writer is not criticizing. Also it might be thought that the expansion of the service
industries contradicts the statements in the text. But a moment's reflection should
convince the reader that the contrary is the case, for certain service industries are nothing
but labor-saving organizations for the performance of services which without them would
require very much more labor. Apartment hotels "with maid and valet service" are an
instance.
1919-1929 843
increase in the total of the national wage bill, have increased total expendi-
ture on commodities, but would only have redistributed it among con-
sumers. This is best seen in the case of that type of employment which
may be designated by the term help: domestic service, shop assistance
in small retail trade and in the business of the carpenter, painter, plumber,
and so on, even in the small and medium-sized factories. For the total
amount of employment these cases, persistently neglected by the current
theories of wages, are extremely important, though relatively more so
in Germany and England than in the United States. And they display
a particularly simple relation between employment and wage rates.
No other application of the traditional schema of economic rationality
is so obviously true to real life as the picture of the owner-manager
of a retail shop who "balances" the advantage of not having to get up
early in order to sweep his shop against the advantage of indulging in
marginal glasses of beer. But although changes in wage rates are
promptly reacted to in those fields and although these reactions are by
no means indifferent for the sum total of satisfactions and incomes,
they do not affect the statistical measures of industrial output or the
revenue of manufacturing or commercial industry. For business at
large it is indifferent whether the retailer hires a help that will buy con-
sumers' goods, or does the spending himself. What happens in most of
these cases is that some earners of what primarily are labor incomes share
their flow of consumers' goods with others in return for the latter's
services. Perhaps another million men and women could have been
inserted into "productive" activity at a comparatively small sacrifice in
rates. And the national wage bill in the statistical sense might have
been increased correspondingly. But the difference this would have
made to the general complexion of business situations would, neverthe-
less, have been small.1
1 Those of us who believe in the stronger saving propensities of the shopkeeper and in
their catastrophic effects on the economic process will have to rate that difference more
highly. Even so, the above argument would hold approximately. There would, in this
case, also be another argument against high- wage theories, which the writer, however,
does not care to stress. We may note in passing still another line of reasoning according
to which a higher or lower level of wage rates acts on employment and output through the
higher or lower rates of interest which it induces not, as older doctrine would have argued,
by relatively decreasing or increasing the supply of "real capital," but by decreasing or
increasing unused lending facilities. It is true that variations in the price of a commodity
quantitatively so important as labor influence all other prices and quantities and also all
monetary magnitudes and expressions. This is, in fact, why a complete theory of wages is
so very complicated a matter. But the simple nexus alluded to can be asserted to be
operative only by means of so unrealistic assumptions and owes the importance attributed
to it so exclusively to a theoretical model which excludes all the vital mechanisms through
which variations in wage rates act that we need not proceed with it. Under its assump-
tions the proposition is, of course, tautologically true.
844 BUSINESS CYCLES
It will be seen that this analysis does not exclude the possibility that
still higher wage rates might have netted a higher total monetary or
even real wage bill. The best method of convincing ourselves of this is to
envisage a comprehensive organization of all employees acting as a dis-
criminating monopolist. This monopolist would have had to go in many
spots below the rates that were actually paid, especially if he had acted
with a view to maximizing real wage bill in the long run. But that both
the weighted index of money-wage rates and the resulting wage bill
might have been higher than they were is not only possible but plausible.
Our argument, together with some of the facts glanced at, no doubt
suggests that in that case total output might have been smaller or, at
all events, not appreciably greater and that there would have been still
more unemployment. But all we have established is that employment
of labor per given amount of product and direct consumption of services
of labor were less than they would have been with relatively lower wage
rates. Although we have also seen that this component of total employ-
ment was as a matter of fact very strong and that any increase in total
output which might have been produced by a further rise in wages would
have had to be very great to counterbalance it, the case is one of rea-
sonable likelihood only. There certainly was no sign of "inadequate"
consumers' spending, whether due to saving (admitting for argument's
sake that saving would have had that effect) or to any other causes, and
there were reasons more convincing than that for such underemploy-
ment of resources as there was. A theory of the world crisis can be no
more derived from any effects of "too-low" than it can from effects of
"too-high" wages and all the attempts in this direction — e.g., consumers'
not being able to buy what was, or would have been, produced — enter
into well-known categories of provable error. But that does not consti-
tute exact proof of the presence or absence of disturbance emanating
from wage rates.1
3. Much of the above applies, mutatis mutandis, also to the German
case (Chart XLVIII).
The wage bill, owing to its comprehensiveness, carries different
meaning, however, and the data about unemployment are more nearly
exact, although it is still necessary to allow for some unseen unemploy-
1 The reader will observe that this is exclusively due to the data of the period which,
apart from their inadequacies, do not allow us to insert into the framework of general
theory, which cannot do more than formulate questions and describe possible cases, factual
assumptions definite enough to enable us to choose between these cases. There are
other situations — we shall meet one in the next chapter — in which it is easier to do this.
Other economists feel no compunction about definite assertions in any case. This is due
partly to the simplicity of the theoretical models which are satisfactory to them, partly
to the readiness to embrace particular factual assumptions.
1919-1929
845
ment,1 and although a change in the attitude of the public mind and the
possibilities of abuse which every system of unemployment insurance
offers may have had something to do with the absolute figures.2 More-
over, it must be remembered — this, of course, also applies to the United
Kingdom — that, while in the United States hourly earnings were (at
least they would be if we imagine them to be calculated with ideal cor-
COST OF LIVING?
WHOLESALE PRICES
1925
1926
1927 1928 1929 1930 1931 1932 1933
CHART XLVIII. — Germany (see Appendix, p. 1071).
1934 1935
rectness) approximately equal to the money cost per hour of employing
labor, in Germany the two differed significantly, especially by employers'
contributions to social insurances.
From January 1925 to December 1929 the Reichsamt figures for
average hourly trade-union rates (tarifliche Stundenlohnsatze) increased
by 47 per cent.3 This increase was associated, on the one hand, with an
1 By unseen unemployment we mean unemployment which escapes statistical measure-
ment, for example, because it outlasts the benefit period. We do not mean unemployment
which, being only the result of artificial definitions, exists but in the pages of some theorists
and may include individuals who think themselves fully employed.
2 There is no doubt that those abuses have been recklessly exaggerated by some and
recklessly denied by others. The war waged by the authorities upon what was termed
black labor indicates, however, that they were not negligible. But we need not believe,
or attach much importance to, the drastic anecdotes that were current about them.
8 They continued to increase in 1930, but not much. It must be noticed, however, that
the "effective" rates (= official basic rates plus various additions, Zuschltige), which- we
846 BUSINESS CYCLES
increase of about 10 per cent in cost of living, and on the other hand,
with a substantial increase in unemployment.1 Neither was simply
"due" to that increase in rates. Concerning the first, there were several
other factors, the most important of which were the agrarian policy
(effective in the last 3 years of the period) and the gradual normaliza-
tion of house rent (expenditure on shelter increased by 55 per cent from
1925 to 1929). But there remains a residue which, being contrary to
Kondratieff expectation, can hardly be attributed to anything else but
to increase in wage rates. This would mean, of course, that the latter
was effective in raising the monetary wage bill above what it would
otherwise have been: for that increase in wage bill which is the "natural"
consequence of the downgrade process would not, as we know from both
theory and history, increase the cost of living. It is, however, perfectly
plausible that in this case and for the time being the increase in rates
actually resulted in a considerable increase in wage bill, because it was
crowded into less than 4 years, which is hardly enough for the system to
adapt itself fully.
Concerning unemployment, the relevance, not only of burdens directly
associated with the employment of labor, but also of burdens apparently
unconnected with it, must again be stressed. We have seen that high
corporation and income taxes increase the sensitiveness of the economic
process to most other disturbances and in particular to any increase in
costs. Since taxation of this type was the outstanding economic feature
of postwar Germany and higher than anywhere else, it is not possible to
disentangle the effects which the increase in wage rates might have had
if it had occurred alone. Perhaps there was some truth in the contention
of the exponents of trade-union interests that the wage rates did not
per se, at least temporarily, substantially contribute to the abnormal
and rising amount of unemployment that prevailed practically through-
out.2 We can certainly not use the reasoning that served in the American
case, for German wage rates started from what even in 1925 was a low
level. But what the high general level kept by wage rates throughout
have for only some industries, rose less. Even if we take "tariff" rates, their increase
does not mean, of course, that even in 1930 they — or earnings — were high according to
general cultural standards. Real hourly rates were only about 10 per cent above 1914 in
1929 and only 16 per cent in 1930. On the other hand, if it is cultural standards and welfare
considerations that we have in mind, the considerable unearned increment in real incomes
must be taken into account which accrued to the working class from public expenditure.
1 The hyperseasonal peak in the winter 1928-1929 is, however, due to the indirect
effects of the labor struggle in the Ruhr district (the direct effects of strikes and lockouts
are excluded). The peak in the winter 1923-1924, when about one-fourth of all members
of trade unions was unemployed, of course, does not count either for our purpose.
2 The long and acrimonious controversy which was at the time waged on the subject
has lost its interest because the arguments used by either party sound hopelessly antiquated
now. Even if they be appraised ex visu of their dates, little can be said for most of them.
1919-1929 847
the period did in this country, was in Germany largely done, though by
a different route, by the high general level of taxation: wage rates,
"social burdens," and taxes taken together may reasonably be held to
explain conditions in the labor market which neither of them could explain
if taken in isolation.1
4. This is not less true for the United Kingdom. Otherwise, the
case is, at least in one important respect, more like the American than
like the German. We find remarkable stability of money wage rates
from the end of 1924 to the beginning of 1928 and then a slight decline,
which was to become somewhat more pronounced in 1930. As in the
United States, this was preceded by a downward revision from the post-
war peak, which extended from January 1921 to the end of 1923, and
by a recovery from that which covered 19242 but was, in contrast to
what happened in America, only slight. Nevertheless, it left money
wages at about 96 per cent above their 1913 figure — already associated
with some supernormal unemployment — a level which had only to be
kept up in order to increase real wages, thanks to the Kondratieff ten-
dency, free trade, and the monetary policy pursued, to nearly 20 per
cent above 1913 at the end of our period (annual average). See Chart
XLIX.
This is what trade unions and the various public agencies3 attempted
and, during our period, achieved. The slight upward pull exerted by
the Juglar prosperity is not visible in the chart and barely so in the
series — the index increased by one point for a short spell (December
1 The above is confined to the years which followed upon inflation, stabilization and its
immediate effects. As mentioned in a previous note, there was "disturbance unemploy-
ment" in the winter 1923-19124. But it is worth noting that the rest of 1924 displayed
only a moderate number of totally unemployed — a fact which was associated with a very
moderate level of rates.
2 There is a statistical difficulty about these statements. We are using Professor
Bowley's index of weekly wages (see Appendix, description of Chart XLIX), which from
January 1925 runs on a new basis. The two series have been "spliced/' but no great
confidence can be placed in this procedure. This is why we refrain from statements about
the course of wages during the war. According to the old series (not shown in the chart,
except as transformed by splicing), rates would have followed cost of living with a lag and
caught up with it in 1919, more definitely in 1920. The maximum of rates (January 1921 ;
277 per cent of 1913) would have followed the maximum of cost of living (278 per cent)
with a lag of three months.
3 Along with the increase in the power of trade unions, public regulation of wages devel-
oped from the Minimum Wage Act of 1909 to the Amending Act of 1918, the Joint Indus-
trial Councils, the Railway Conciliation Councils, and the Agricultural Wage Boards. If
we include the parallel activities of trade unions and take account of the fact that official
awards will exert influence beyond the cases decided and beyond the trades under jurisdic-
tion, we may say that almost all wage contracts in the United Kingdom are publicly
controlled. This, of course, raises a problem in the interpretation of our curves. But it is
believed that what follows in the text is not open to objection on this ground.
848
BUSINESS CYCLES
1926 to April 1927) — and the unusual lack of covariation with profits
which have been inserted in this chart in order to display it, is significant.
These features as well as the attendant unemployment, no doubt, invite
an interpretation similar to that we have tentatively adopted in the
American case. This — i.e., that the general level of wages was "too
high" from the start — was, in fact, the opinion of most English econ-
omists,1 and an impression to that effect must have prevailed among
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART XLIX.— United Kingdom (see Appendix, p. 1072).
trade-union leaders, who, though they energetically defended that level,
never seriously tried to raise it. The present writer doubts its correct-
ness, however. That increase in real wages was not so very impressive.
It was certainly smaller than it would have been without the war and
hardly greater than we should expect it to be in spite of the war, during a
Kondratieff downgrade and in a country that as yet substantially
adhered to free trade. Any explanation of strains and maladjustments
that rests on an increase of real wages by, from 1924 to 1929, little more
1 That fact is veiled not only by the guardedness of the statements of some of them
but also by the fact that the concurrence of others was expressed in an indirect manner,
which is likely to escape notice. Advocacy of "monetary expansion" may, advocacy of
protective tariffs must, imply attack on real wages while it is at the same time natural for
the advocate of these measures to oppose a reduction in money wages.
1919-1929 849
than 1 per cent per year seems hazardous, if not downright misleading.1
The national wage bill became burdensome because it was accompanied
by a fiscal policy that made it so. If the labor interest did not press for
an increase in wage rates, it attained the same object by pressing for,
or putting out of court the reform of, taxation which effected a transfer of
wealth2 much greater than any conceivable increase in wage rates could
have effected.
Many other factors, however, swelled unemployment percentages
during the period under discussion. The great problem that attracted
so much anxious attention dates from 1923. Until the last quarter of
1920 the trade-union percentage (unemployed males) was understand-
ably below anything that can be called normal according to any stand-
ard. It soared to over 23 per cent in June 1921, but this, too, was readily
understandable. In 1923, however — we now shift to the percentage of
unemployed insured persons, which was (annual average) 17 in 1921 and
14.3 in 1922 — unemployment was still at 11.7 per cent and in the vicinity
of that figure it stayed3 throughout (1924, 10.3; 1925, 11.3; 1926, in spite
of strikers disqualified for benefit being excluded,4 12.5; 1927, 9.7; 1928,
10.8; 1929, 10.4). There is, as has been stated before, no significant
"trend" in this but, even apart from the figure for 1926, which reflects
the indirect influences of the great struggle of that year, a "level" of
about 10 per cent. Now this cannot be compared to prewar (trade-union)
percentages, because the percentage of total unemployment is, of course,
likely to be greater than the trade-union percentage which served as
indicator in the prewar epoch; because, for the same reasons as in Ger-
many, statistically visible unemployment would presumably be greater
in the postwar period, even if we had otherwise comparable figures;
and because trade-union regulation, social insurance, and other factors
made for reduced geographical and industrial mobility of labor. For
these reasons normal unemployment in our sense of the term would,
ceteris paribus, be greater than it had been. It has been estimated that
even in times of active business it would not fall below from 2 to 4 per
cent,5 and we will tentatively accept 3 per cent as a compromise. Second,
1 The above statement is intended to be read with previous discussions in mind that
cannot be repeated. Unconnected with these, it would, of course, be meaningless. It
should be added that that increase in real wages was largely at the expense of foreign
producers.
2 The reader will observe that the above statement stresses taxation only and not also
expenditure benefiting the masses. For such as it actually was, this would have been
possible with almost no interference with the efficiency of the capitalist machine.
3 Trade-union figures were, however, materially lower in 1924 (minimum of 7 per cent
in May).
4 So they were in 1921.
* See Colin Clark, Statistical Studies, Economic Journal for September 1931, p. 849.
The trade-union percentage was, however, as low as 0.9 in April 1920.
850 BUSINESS CYCLES
there is the normal effect of the Kondratieff downgrade to take account
of, which may, according to the experience of the seventies and eighties,
easily double that figure. This already includes the industrial and
commercial shifts and readjustments incident to the downgrade processes,
but only so much of them as the internal evolution of the country would
entail. The great shift from world market to home production which
English industry had to undergo in consequence of the shrinkage of her
exports is, third, not covered by that estimate. If we accept Mr. Colin
Clark's figure1 that, if England's "position as an exporting co*untry had
not been deteriorating, unemployment in 1929 would have averaged
about 900,000 as against 1,250,000," and neglect the possibility that
this deterioration may also have had something to do with wages, we
have here an independent factor the influence of which we will put at
2 per cent in the average, which is certainly conservative. Thus we
arrive at the result that 8 out of the 10 per cent unemployment level of
the period can plausibly be accounted for, if not entirely without reference
to labor policy, yet without reference to any behavior of wage rates other
than what it would have been within the most normal working of our
process and, substantially, even under laissez-faire conditions. This
statement is open to objection on statistical and theoretical grounds;2
but it still serves, so the writer believes, to give a rough idea of the order
of magnitude of what may remain for explanation — as vicarious unem-
ployment— by "rigid" wage rates plus fiscal policy. However that may
be, it cannot well be doubted that the postwar unemployment problem
was largely one, not of systematic tendencies, but of disturbance by
external factors.
III. In proceeding to discuss the course of events in the banking and
cognate spheres we will remind ourselves again that this sequence of
topics is not to suggest a progress from effects to causes and that, though
there is nothing but interdependence between the quantities, monetary
and other, which enter into our process, we should, were we constrained
to set up causal nexus at all, prefer to put our trust in the reverse one.
1 Op. cit., p. 349. See ante, Sec. E, 1.
2 The indubitable truth, stressed by Professor Pigou in his Theory of Unemployment,
that it is impossible to distribute a given amount of unemployment among different causes,
does not, however, stand in our way, because there is no objection to trying to estimate
the difference which presence or absence of a given factor makes or would make and it is
in this sense that our statement should be understood. But it is more serious that the
other factors would of course have produced different results at different wage levels and
can never be treated independently of them. We restate, therefore: Assuming that, given
the fiscal and monetary policies actually pursued, the policy of trade unions (including the
threat of general unrest) and of the public wage-fixing agencies kept money wage rates above
what they would have been without those unions and agencies, then, whatever that dif-
ference was, its responsibility for unemployment can at most have amounted to 2 per cent
of the total of insured persons.
1919-1929 851
This reminder is justified by the fact that it was primarily with reference
to postwar developments and in connection with postwar problems that
the "deposit logic" committed its most flagrant excesses. It should
also be borne in mind that in no country, but especially not in the United
States, did banking figures retain their prewar significance — not so much
because of statistical reasons as because of fundamental institutional
changes: all the more important is it to see the old essence under new
forms and phraseologies where the old essence did persist.1 Our dis-
cussion will almost wholly be confined to the American developments.
a. We begin by clearing up a point which may have puzzled the
reader in inspecting Chart XL VI. While in the United States postwar
pulse chart balances have been represented, in deference to prevailing
opinion, by outside net demand deposits plus "circulation" (for explana-
tion, see Appendix), debits have been compared on Chart XL VI with
total outside deposits (demand plus time) minus outside investments.
According to that prevailing opinion, only demand deposits are balances
in our sense or, as most authors prefer to put it, "money."2 We have a
good series for what is referred to as adjusted demand deposits, which is
cleared of interbank and includes government deposits,3 and most of us
use net demand deposits as a substitute for purposes for which the other
series is not available. But time deposits are held to have no more claim
to be included than bonds would have were banks in the habit of issuing
them. This view seems correct as far as time deposits represent house-
hold investments — genuine "saving deposits" — but for the following
reasons it does not seem to be so with respect to the bulk of time deposits
in commercial banks — and a great part of the time deposits in savings
institutions — which does not constitute investment and has nothing to
do with savings and the spectacular increase in which it is a mistake to
treat as an indication of changes in the rate of saving.4
1 It may also be mentioned that the figures of national banks, notwithstanding the
changes made in their position by the Federal Reserve Act and the fact that their propor-
tionate importance has been on the decrease, might still be used for our period with approxi-
mately as much justification as there was in the prewar time. The late A. A. Young was
of the same opinion. We shall, with one exception, not go into seasonal variations, which
have been considerably affected by the institutional changes, and will hence mention here
that the autumnal drain of cash from New York had lost much of its importance. On
seasonal variations in general, see A. A. Young, Analysis of Bank Statistics, p. 53 et aeq.
2 We do not now follow that usance in order to avoid a needless controversy. The
term Means of Payment is less open to objection, but may still include things that nobody
wishes to include, owing to the ambiguity of the word payment.
3 See L. Currie, The Supply and Control of Money, 1934, p. 13. Study of Chap. Ill
of that book is recommended as an introduction into the statistical difficulties of the
subject.
4 The exaggerated ideas some writers entertain about postwar saving activity are
sometimes precisely due to that mistake. It has been pointed out before, for example,
852 BUSINESS CYCLES
First, as far as time deposits were actually drawn against, they served
from the standpoint of holders exactly the same purpose as demand
deposits.1 Opinions differ as to the extent of the practice (which was
prohibited in 1933), but its importance cannot be read off from statistics
of turnover. This, of course, is much smaller than that of demand
deposits. But on the one hand, it is naturally slow-moving "cash" that
is held on time deposits and, on the other hand, it is the possibility of
drawing (practically)2 at will that matters. Second, even if no check
had ever been drawn against time deposits and if, as in the case* of bonds,
one would really have had to "convert them into money," they would
still have been so very like cash — since, unlike bonds, they can always be
turned into cash at par — that the making of the distinction comes, from
the standpoint of the holder, very near to hairsplitting. Whoever holds
that kind of asset will behave differently with respect to his demand
balances, in particular, feel much less constrained than he otherwise
would to keep an emergency reserve. By the classing of time deposits
with customers' investments an important feature of the monetary situ-
ation is completely lost.3
Third, a growing habit of keeping on time account as much of one's
cash balance as is possible or convenient would suffice to explain the
growth of time deposits relatively to demand deposits. Suppose that
that Mr. Lough's estimates of savings are partly due to his treatment of time deposits,
although he is not unaware of the point and tries to meet it.
1 Dr. Currie, op. cit.t p. 15, argues ably against this view. But even if the present
writer had more confidence in the realistic virtues of Dr. Currie' s assumptions, that argu-
ment would not meet the point as formulated above.
2 A bank that refused to honor a check covered by a time deposit, had to be prepared to
face an unpleasant discussion and to lose the customer. The writer in some cases observed
that clerical staffs had blanket power to honor such checks but had to refer to a responsible
executive if they wanted to refuse. Of decisive importance, finally, are those agreements
that seem to have been fairly frequent, especially in the West — so the writer has been
told by a competent authority — according to which holders of time deposits were accorded
a limited right to draw without notice, two or three times a year for example. This is
proof that time deposits were by those banks looked upon as a special kind of demand
deposit to which time-deposit privileges in the shape of higher interest were granted under
the pressure of competition. The Report of the Committee on Member Bank Reserves,
1931, seems to take much the same view. Mention is due to Dr. B. Anderson's important
contributions to the subject of time deposits, see, in particular, Bank Expansion versus
Savings, Chase Economic Bulletin, June 25, 1928.
8 It might be urged that a similar consideration applies to all highly liquid assets. This
is quite true to the extent that neglect of any type of what we have called "near-money"
is, in fact, a bar to correct analysis of monetary processes and a cause of the inadequacy of
the picture drawn by, as well as of the recommendations of, the modern exponents of the
quantity theory or, to use Dr. Anderson's phrase, those economists who know nothing
but "the monotonous tit-tat-to — money, credit, prices." But we might reply that such
assets cannot be normally relied on to represent so prompt a control over a definite sum
as do time deposits.
1919-1929 853
there are in a country only demand and "genuine savings deposits/*
The former all "circulate" but display different rates of activity ranging
from several hundred turnovers to almost zero per year. Then let a new
name, say time accounts, and some privilege be introduced for balances
which are being used less often than an arbitrary number of times per
arbitrary period, people being invited to register for these time accounts
on condition that they use them only at certain dates or at a certain
notice. In the name of all that we know about the behavior of people,
we must assume that they actually will register, first and foremost, with
respect to that part of their demand deposits which they would anyhow
not use before the assigned dates or need not use without notice, and
second, with respect to another part which can be made to qualify with
but little inconvenience. Nothing whatever has changed in the behavior
of those who register. In particular, the holders of these new time accounts
spend exactly or almost exactly what and when they used to spend before.
Yet time deposits have emerged which, incidentally, may reflect credit
creation just as much as demand deposits.
This paradigma is merely intended to illustrate a principle. It is
not held that the new practice will not make any difference in other
respects: if, in particular, legislation imposes higher reserve proportions
on demand than on time deposits, the shift of balances from the former
into the latter category will increase the lending power of banks, which
is precisely how the growth during the twenties of time deposits in this
country became an instrument of monetary expansion; and any payment
out of time deposits, no matter whether effected directly or by means of
a previous conversion into demand deposits, will, if it induces a more
than momentary shift from time to demand account, have the opposite
effect,1 the same as an increase in reserve requirements would have.
1 Dr. Currie rightly emphasizes this. If the bank in question happened to be "loaned
up," it would, in consequence of such a payment, have to reduce its loans or investments
by a corresponding though not equal amount. But if time deposits are freely flowing in
or demand deposits have a tendency to shift to time account — an essentially temporary
state of things, no doubt, but one which prevailed during the twenties — the bank will
in the typical case be relieved of that necessity, even if it were so perfectly loaned up as
Dr. Currie must, by virtue of his theoretical scheme, assume any bank on principle to be
and even if hence any use made of time deposits would necessarily entail some contraction
in other demand deposits. Perhaps another comment will be useful, though it is really
implied in what has been said in the text. Let us return to the hypothesis that payment
can be effected only by draft on demand deposits so that time deposits can serve the
purpose in no other than the indirect way, i.e., through previous acquisition of a demand
deposit. Then firms and households will for each day acquire, against time deposits,
demand deposits sufficient to provide for the expected payments of that day. But not
more. The rest of the balance will stay on time account. Demand deposits and bank
reserves will be "economized" thereby. But exactly as, ex visu of any given day, we do
not confine the concept of deposits to that amount of demand deposits which changes
854 BUSINESS CYCLES
This is why we emphasized the growth or spread of the habit to insure
that there should always be a net shift toward time account. We may
further approximate the paradigma to American reality by granting
that many big concerns and rich households did not stoop to availing
themselves of the advantages, in their nature modest, which we suppose
to have been attached to time accounts,1 so that it may be primarily the
middle and lower strata of depositors who did so; by allowing for savings
institutions' trying to acquire slow-moving balances, once the trick has
been learned, while commercial banks had an additional motive to hunt
for genuine saving deposits, as soon as they were in turn privileged in
handling them;2 and finally, by adding that temporarily idle funds will
naturally tend to be placed on time account.3
hands on that day but extend it for obvious reasons to the total of demand deposits in
existence, so we would for the same reasons have to include the time deposits that are not
converted into demand deposits and hence remain time deposits, even if that part which
is converted were always, necessarily and entirely, balanced by a contingent contraction
of other demand deposits. This is what our paradigma is intended to show. The impor-
tant but nonessential difference in reserve requirements has certain consequences, but it
does not alter the nature of time deposits.
1 Statements about banking practice differ on this point. Dr. Anderson holds that
the time deposits of New York City banks chiefly consisted "in temporarily idle funds of
great business corporations or of foreign banks or of rich investors who had temporarily
disposed of investments and were awaiting opportunities in the market." (A Critical
Analysis of the Book by Lauchlin Currie, address before the New York chapter of the
American Statistical Association, Apr. 26, 1935, p. 20) while others hold that "bank
statistics show clearly that there could have been no appreciable amount of shifting of
deposits from the demand to the time category on the part of large depositors in national
banks" (D. R. French, Significance of Time Deposits in the Expansion of Credit, 1922-1928,
Journal of Political Economy for December 1931). The two statements do not conflict
as much as they seem to. But it is not unlikely that those big concerns which had such
good opportunity for temporary investment resorted less to time accounts than did smaller
ones that had not. And it is just as well to note that this point is not essential to our
argument.
2 Of course, both savings institutions and commercial banks always did do that, as we
have seen before, in this country as well as in others. The Federal Reserve Act only added
another stimulus and, for national banks, additional powers.
3 We hold, agreeing in this with Dr. Anderson, that the bulk of time deposits has much
more to do with credit creation than with savings. This is not refuted by the fact that
deposits evidenced by passbooks averaged almost 72 per cent of total time deposits in
national banks since the autumn of 1928, from which Mr. D. M. Dailey (National Banks
in the Savings Deposits Field, Journal of Business for January 1931) infers that "in excess
of four-fifths of the time deposits reported by the national banks represent the accumula-
tion of savings:" there is no reason for inferring that, though there is for taking the large
use made of the passbook method as an indication that it was, in fact, firms and households
of moderate means which primarily availed themselves of the opportunity offered by time
accounts. But while denying the alleged relation of time deposits to saving, we do not
entirely deny their relation to underspending. In order to bring out what seems to the
writer an important principle, our paradigma has been so framed as to be independent of
1919-1929 855
Fourth, a motivating privilege such as has just been envisaged, was
provided by the reduction made by the Federal Reserve Act in the
amount of reserves to be held against time deposits and the further
reduction enacted in 1917. The fact that the spectacular growth of
time deposits dates from that time strongly suggests a connection.
Although the circumstance that banks offered a higher rate of interest
on time than on demand deposits should suffice to prove that they thought
the former more advantageous than the latter, many writers have argued
that the gain which banks could expect from a transfer of demand depos-
its to time account was, owing to the incident interest charge, small and
that there was, hence, little if any motive for them to encourage their
customers to effect such transfers. That depends on what they did with
the resulting increment in their lending power. Economists who assume
that they would buy governments had no difficulty in showing that the
gain was small. In fact, the writer believes that in some cases there must
have been a trifling loss. Smallness of gain, however, is frequent in the
banking business and many a small loss is undergone in order to acquire
or satisfy an otherwise valuable customer. And if the increase in lending
power went into customers' lines of credit or into mortgages — real estate
loans rose from less than 600 millions in 1919 to nearly 3.2 billions in
1929 (all member banks), a rate of increase much greater than that of
total earning assets — there was not only gain but a very attractive one.
But this is not the point. It is quite sufficient that there was motive
on the part of customers. About that there cannot be any doubt.
Moreover, while a nonnegligible premium on time deposits was offered
variations in the rate of spending and as to show that spectacular growth of time deposits
could come about without any underspending as well as without any saving. It follows
that it would be dangerous to rely on time deposits as an index of people's aversion to
spending. Still, although a demand deposit may, and often does, become as idle as
any time deposit, it is reasonable to assume that sums which are not expected to be wanted
at all for a time are even more likely to migrate to time account than others. This
fits in very well with the violent fluctuations in the rate of growth of total time depos-
its, which to the writer's astonishment Dr. Currie flatly denies (op. cit., p. 132). They
increased from December 1921 to December 1927 (reporting member banks; monthly
averages of weekly figures) by about $3.4 billions. Nearly two-thirds of this increase
occurred from 1921 to 1922, from 1923 to 1924 and from 1926 to 1927 in obvious connection
with banks' investments (see Chart L), which in turn were no less obviously related to
open-market purchases of the Federal system. These two-thirds, at all events, are pure
creations of monetary management and emerged on time account because, in accordance with
our contention, emphasized throughout this book, such creations tend to become idle.
Deposits of all national savings banks, on the contrary, increased in fact much more
steadily (by about 500 millions a year from 1921 to 1927, figures as of June 30) and are
"not affected to any great extent by inflows and outflows of reserve funds," which Dr.
Currie avers with respect to time deposits (p. 99). But for the reason mentioned in the
preceding note we hesitate to accept their rate of increase as a standard by which to dis-
tinguish genuine savings from other time deposits.
856 BUSINESS CYCLES
to them in the shape of a rate of interest considerably higher than that
which was paid on demand deposits, changing habits greatly reduced
the incident inconvenience, at least for households: the greater the rela-
tive importance in the budget of expenditure on durable goods, payment
for which is lumpy and postponable, and the more widespread the prac-
tice of charge accounts, the less need there is for absolutely ready cash or
advantage in holding it.
The view that in the twenties time and demand deposits were essen-
tially the same kind of thing is verified by the relation to each other of
the two top curves on Chart XL VI. From total outside deposits, banks'
investments have been deducted on the theory that they tend to be
idle. In doing this we are no doubt overshooting the mark — for we know
that even for prewar times it would not be true to hold that banks'
investments always proceed from the banks' initiative — as we are over-
shooting it by including all time deposits. All the more interesting is
the result of the experiment. Our deposits curve fits the debits curve
excellently from 1921 to 1927 and the different behavior of the two curves
in 1919 and 1920 and again in 1928 and 1929 is easily accounted for by
the irregularities of those years, mainly by the excesses of speculation,
which would also influence outside debits. The curve of outside demand
deposits does not fit so well and thereby raises a typical spurious problem,
viz., why the "velocity" of these deposits should have so much increased1
and how it was possible for it to do so or, to put the matter differently,
how it was that outside debits behaved as they did (increasing from 17.6
billions in 1919 to 27.7 in 1929) with demand deposits increasing only
from a little over 6 billions to about 8.2 The much more natural complex-
ion of the curve of "velocity" of total outside deposits (see again Chart
XLVI) will strengthen our belief that we have in fact got hold of the
approximately true figure and that our diagnosis of the nature of time
deposits is correct.3 But it must be added that there is another possible
1 The index of the rate of turnover of demand deposits in principal cities, which has
been constructed by the Federal Reserve Bank of New York (1919-1925 = 100), in fact
starts with 100 at the beginning of 1921 and after a fall returns to it in 1922, then remains
at that level through 1924, after which it rises, with an interruption in 1926, until it reaches
200 in 1929. This includes New York. But if we exclude New York, the paradox remains,
although toned down.
2 Monthly average of deposits other than U. S. Government deposits, including Due to
minus (Due from plus Items in Process of Collection). Dr. Currie's series of total money
"supply" (op. cit., p. 83; including Cash outside Banks; objections to the term Supply,
which in this connection is so pregnant with misleading associations, need not be repeated)
displays, of course, the same phenomenon. It increases from nearly 22 billions in 1921
(when total debits were nearly 88.3 billions) to nearly 26.7 in 1929 (when toted debits were
nearly 78 billions). Since total debits are heavily loaded with speculative transactions,
this is not more but less striking than the comparison in the text.
8 It may be said at once that the same diagnosis applies to English deposit accounts and
to what corresponds to these in Germany although, owing to the absence of the particular
incentive imparted by American banking legislation, to a lesser extent.
1919-1929 857
reason why comparatively so small an increase in outside demand deposits
should have proved adequate. War and postwar boom profits were not
annihilated by the subsequent crash. Only an insignificant shrinkage of
outside demand deposits occurred in 1921. There is a difference of only
about 1 billion between the monthly peak in 1920 and the monthly trough
in 1921 and this was quickly made up. In other words, we see again that
the monetary system of the United States stayed expanded after the war
and did not contract in response to the temporary downward revision of
values. Thus, the latter left people amply provided with owned funds
and for a time at least, i.e., while this slack was being taken up, the eco-
nomic body was to some extent able to grow within the existing coat.1
However, taken at its highest possible value, this explanation cannot be
considered as sufficient, even to the end of 1924; while after that, i.e.*
precisely for the years when the phenomenon of increasing "velocity" is
most clearly marked, it is not available at all.
b. Total deposits minus investment, then, unlike demand deposits,
display during the period the general tendency or descriptive trend that
prewar experience leads us to expect, though the "normality" of the
observed gradient, in this case as it was in that of wage rates, is compat-
ible with an abnormally high level owing to the location of the starting
point. Accepting the latter, however, we also see that fluctuations
themselves were not less normal from our standpoint than the descriptive
trend they produced. In particular, we see the footprints of the Juglar
recovery and the four phases of the last Kitchin of the third Juglar.2 In
1925 and after, the prosperity and recession of the fourth Juglar are well
marked by an increase in our quantity and by its tapering off. While
this does not show in outside demand deposits alone, another feature of
special interest can be observed just as well in these: the behavior
of deposits with respect to debits, from the last quarter of 1929 on.
While debits fell sharply, both outside demand deposits and total outside
1 Some part of the surplus funds which were set free by the reduction of operations
and the fall in prices may have migrated to time account very early in the downturn, or
even before. This possibly explains part of the strong increase in (total) time deposits in
1920, the monthly average of which year is nearly 800 millions above that of 1919. The
general swerve toward the time account must, however, not be forgotten and makes it
difficult to be very positive on the point.
2 These we also see in the demand deposits, which do not otherwise reflect well either
the "trend" or the fluctuations. It has been pointed out, in proof of a fundamental change
from prewar conditions and of the dominating influence of open-market operations that
demand deposits increased strongly in the years of relapse, 1924 and 1927. Without
denying the element of truth in this — even our series is, of course, not quite independent of
open-market operations, although much less affected by them — it should be observed
that the deviation from prewar experience is so very obvious only in yearly figures and total
demand deposits. As to the case of 1924, for instance, even total demand deposits reflect
the business situation quite faithfully by sagging from the middle of 1928 to the middle of
1924 in correspondence to loans and discounts (see below). And New York demand depos-
its are a different matter and should not be included.
858
BUSINESS CYCLES
deposits minus investments fell much less. As we saw at the beginning
of the period that a cloud of potential inflation failed to burst, so we see
now that in 1930 business operations contracted in the midst of a most
plentiful "supply" of money — the coat refusing to shrink in proportion
to the shrinking of the body. The behavior of total outside deposits,
total deposits, and New York net demand deposits, and the relations of
these scries to that of outside debits should now be studied (Chart L).
NET DEMAND DEPOSITS
NEW YORK CITY
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 193! 1932 1933 1934
CHART L. — United States (see Appendix, p. 1072).
The outstanding fact of very close covariation between Outside Loans
(and Discounts) and Outside Debits next calls for our attention. Clearly
the former dominated the latter during our period quite as much and in
quite the same sense as ever. This emphatically normal state of things
loses nothing of its importance by two features which have been much
stressed in the discussion of postwar banking figures.
First, it will be observed that while dominating debits, loans did
not dominate the link between the two, viz., deposits. This is obvious
in the case of net demand deposits, though not so much as some theorists
1919-1929 859
would have us believe,1 but to a lesser extent it is also true of time plus
demand deposits. It is not true only for our quantity, for which covaria-
tion with loans is, of course, tautological. But this circumstance testi-
fies precisely for our view and against the revised quantity theories, for
it shows that, broadly, only those variations in deposits which corre-
sponded to loans, turned into variations of expenditure, and that those
which owed their existence to other factors did not, whereas according
to those theories they should have done so just as much as any other.
Once more, facts conspire to suggest that regulating deposits does not
imply regulating expenditure or the pulse of business. And this means
that the new feature in the behavior of deposits and the factors respon-
sible for it cannot claim the significance widely attributed to them,
though, as we shall see, they may claim another.
Second, it has often been pointed out that, whatever total loans and
discounts may have done, the commercial loans as represented by the
All-Other category of banking statistics2 remained almost stationary
throughout the period. It is true that if we pass by the first years of the
period and take for a starting point 1923, when liquidation of the crisis
may be assumed to have been over, we get an almost steady upward
movement and a total increase amounting by 1929 to nearly 2 billions
for weekly reporting member banks in principal cities.3 But if we
exclude real estate loans, the figure for all member banks (June 30)
never reaches the 1921 peak, and only 1926, 1928, and 1929 display any
marked increase. Even if we do not do this, the share of all other in
total loans declined from about 70 per cent in 1921 to below 55 per cent
in 1929. This, of course, is no new tendency but only continues one that
was present (in national bank figures) before the war. It is new only in
that it now spread to absolute figures.4 If, however, only quantitative
1 A. A. Young (op. cit., p. 62) rightly emphasizes, for national banks, "the general
correlation of the movements of loans and discounts and of demand deposits both in and
outside of New York City." The divergence, however, asserted itself mainly after the
date (or from shortly before the date) at which his discussion ends. And there was, of
course, considerable divergence in details. The more important cases are associated
with the investment item. But this explanation is not always sufficient and in some cases
nice problems arise into which, however, it is not possible to enter here.
2 Let us recall that the best that can be said for that category is that it has more to do
with what is usually referred to as commercial credit than others. But many loans on
securities are no less genuinely business loans. The greater part of loans on real estate
is not.
3 See the Monthly Review of the Federal Reserve Bank of New York for Oct. 1, 1932, p.
78. New York is included. Professor Mills (op cit., p. 450) gives, for all reporting member
banks, also including New York, an average annual rate of increase of 8.2 per cent, for
loans on securities 10 5 per cent, between 1922 and 1929.
4 Dr. Anderson (Chase Economic Bulletin for Apr. 8, 1927, p. 18) drew attention to the
low percentage of total member bank assets formed by paper eligible in the technical sense.
Dr. Currie also stresses this point (op. cit., p. 117). The theoretical schema underlying the
860 BUSINESS CYCLES
and not qualitative, the change is, within the secondary order of impor-
tance that attaches to the figures and processes of the sphere of credit, a
fundamental one and, hence, requires careful consideration.
One thing is clear: it did not originate with banks since these stand to
lose by it what, in the long run, is their most profitable line of business.
It must have been the customers who turned away from this method of
financing. Among them we can discard the households, for these con-
tinued to borrow freely — in fact, more freely than ever — as the develop-
ment of real estate loans and of personal loan departments suflacesto show.
They also borrowed indirectly from banks, through various types of
financing agencies which discounted with banks, and through helping
to create installment paper, which also formed an increasing percentage
of all other loans. Nor can the small and medium-sized trading and
manufacturing business be responsible for the change, for this remained,
much as it had always been, dependent upon the customers' line of credit
for all its needs, the financing of essentially long-term investments
included. Corporate business, hence, is what, in accordance with
common knowledge and general opinion, emerges from this process of
elimination. But why should bigger and big business have taken this
new departure? Simply because the plethora of money made it easy
and profitable to embark upon a course which in itself appeals to execu-
tives who are always jealous of anything that involves a certain amount
of supervision and who for this reason never love their banking connec-
tions. As stated above, in the first years of the period the more successful
concerns had at their disposal ample funds with which to finance them-
selves. These had been previously assembled from profits1 and in many
cases preserved by timely withdrawal from the firing line. Later on,
when bond yields declined and stock markets boomed, money flowed so
easily toward corporate industry, at rates with which no bank could
compete that, taking full advantage of this situation, concerns became,
in the process of expansion, creditors rather than debtors as far as ready
money is concerned, keeping financially ahead of requirements and
eventually entering the great depression with a financial outfit which was
nothing short of luxurious.2 Some of them were even able to finance
Federal Reserve Act increasingly failed to fit American conditions. The former author
dislikes, the latter likes this development, but for us it is only the facts that matter.
1 We may recall that undivided profits of corporations, unconnected for difference
between depreciation at cost and at current prices, amounted to 4,310 millions in 1919, a
figure never reached again. But financing from profits of course continued.
2 The cash item of corporate balance sheets reflects this imperfectly, but it does reflect it.
Its percentage relation to total assets can be calculated from the corporation tax material,
from 1926 on. Including with cash tax-exempt securities, which served as temporary
investments, and deducting from this item for All Corporations the cash plus tax-exempt
investments held by banks and financial corporations, we get for industrial and commercial
1919-1929 861
their investments in 1935 and 1936 by funds that had been raised during
the speculative mania of 1928 and 1929, 1
Now, that " money '* still came from or come through banks. It was
to a great extent created by them, exactly as it would have been had
corporations directly borrowed from them. Only it was not created
through loans to industry, which would have swelled other loans, but
through loans to buyers of bonds and stocks, which swelled loans on
securities. At least this is that alternative to lending directly to indus-
try which will bring out most clearly the point we are trying to make.
Banks, of course, also bought bonds — and stocks through their affiliates —
and placing these acquisitions became so important that a new type of
bank executive emerged who had little of the banker and looked much
like a bond salesman. But this we will put aside for the moment in order
to focus attention on the fact that to a considerable extent credit creation
by loans on securities was a substitute for "business loans" — hence, to
the same extent not a net addition to total volume of balances — and,
inasmuch as it was, served the same purpose and not an additional one.
It does not follow that the actual amount of balances created was the
same as that which would have been created by the orthodox method
or that the change, being one of mere technique, was unimportant. On
the contrary, it is safe to say that in the midst of rioting stock markets
creation went on at a pace very different from that which would have
been set in bankers' conference rooms. The steering and balancing parts
of the capitalist machine were seriously and perhaps permanently
impaired. From remedies eventually applied to the ensuing situation a
novel apparatus of banking may well evolve from which there is no road
back to what, from the standpoint of our model, would have to be called
normal conditions. Nonetheless it is necessary to take account of that
substitution of one method of deposit manufacture for another if we wish
to speak — the writer does not care whether we do or not — of the "infla-
tion" of the twenties. And for the immediate purpose in hand it is
corporations the following amounts in billions (we add the years of the great depression) :
1926, 9.9; 1927, 10.3; 1928, 11.1; 1929, 10.9; 1930, 10.4; 1931, 9.1; 1932, 9.1; 1933, 8.8; or
in percentages of total assets: 5.65, 5.86, 6.13, 5.57, 5.37, 5.20, 5.37, 5.41 per cent — a series
the remarkable stability of which is well worth noticing.
1 Acceptances and commercial paper bought by banks are, to September 1934, included
in Other Loans. But from the standpoint of the relations between corporate industry
and banks, these items should also be excluded and listed among the methods of inde-
pendent financing. For big corporations either directly or — through intermediate agencies
owned by them, such as acceptance corporations — indirectly appealed to the open market
much as the Federal Treasury did. And although banks were faute de mieux forced to
buy such paper, these purchases do not constitute " discounting for customers." This
practice, in fact, meant, in the short-term field, as complete a rupture of the old bank-
customer relation as bond issues meant in the long-loan field.
862 BUSINESS CYCLES
important to recognize the same old process of financing enterprise in
the new garb, the material of which includes the security loans, and to
see both the stagnation of "commercial" and the expansion of security
loans in this light.
Total outside loans and discounts, then, are the really relevant figure.
Their share in total earning assets varied from 1922 to 1928 between 71
and 74 per cent.1 And there is nothing obviously abnormal or unprece-
dented in their average rate of increase — their "descriptive trend" —
during the period. They also move in the Kitchins, and show the rise
of the fourth Juglar, according to expectation. Divergences in the
behavior of the components give rise to various problems of detail into
which we cannot enter. First differences of the original items of the
all-other loan series reflect short-run business situations more faithfully
than first differences of the original items of the security loan series —
which is easy to understand on grounds of financial technique — so that
the former series remains, for some purposes, more useful than the latter.
c. Turning now to members' investments,2 we will avail ourselves
of the opportunity to look at some of the developments discussed under
(a) and (b) from another standpoint, viz., the standpoint of the banks.
Parallelism between loans and deposits has, of course, never been perfect
but always interfered with by, first, the efflux from and the influx into
banks of currency; second, the efflux from and the influx into the country
of the monetary metal (including efflux into and influx from the arts) ;
and, third, banks' investments. During the period under discussion,
none of these factors operated as it did before the war. But the change
that primarily impressed theorists and, in fact, amounted to a revolution
in banking practice and in the structure of earning assets occurred in the
absolute and relative importance of the third (see Chart L).
This change dates far back. As we know, it is clearly indicated in
national bank figures ever since the nineties, and that theoretical inter-
pretation to which we have grown accustomed — the investment theory
of banking — could have been inspired nearly as well by the course of
events from 1899 to 1908, when "secured" loans and investments behaved
1 1920, 1921, and 1929 display higher pecentages. It is of some interest to note that
there is a rough covariation between the year-to-year changes in these percentages and the
"velocity" of deposits; between the ratio of loans and discounts to net demand deposits
and the commercial paper rate; and between the latter and the ratio of loans to banks'
investments. Interpretation from our standpoint is too obvious to detain us.
2 The figures plotted on Chart L are, in fact, those of (weekly reporting) member banks
outside New York (see Appendix). But the reader should remember that within our
general reason' ng the term Member Bank or simply Member has a technical meaning of
our own. Weekly reporting member banks outside of New York are, on the one hand,
only a sample of a more comprehensive universe designated by this term in its usual
meaning and for us, on the other hand, only the real-world representatives of a theoretical
entity (= all noncentral banks).
1919-1929 863
in a manner highly suggestive of postwar developments. The Federal
Reserve Act and the amendment of 1917, on the one hand, and the gold
influx, on the other, providing additional facilities, war finance imparted
an impulse1 which would not in itself have meant more than a very
natural temporary deviation, under the pressure of exceptional circum-
stances, of no permanent importance for the financial mechanism of
cycles. We observe, in fact, that as soon as those circumstances had
passed, banks were anxious to normalize their position and to reduce their
share in the iceberg that was swimming about and encumbering naviga-
tion in the money market. Since this was done during 1919 and 1920, 2
when the postwar boom was in full swing, we may infer that, while
maneuvering back to normal conditions, they were also maneuvering
for room to satisfy industrial and commercial requirements.3 But this
endeavor to eliminate what, at the beginning of the period, they (and
also the Federal Reserve Board) seem to have felt to be an abnormal and
undesirable situation, was soon given up and never resumed. By the
end of 1922 total investments were almost back to the peak figure,4 to
increase to ever new levels for the rest of the period.6
Interpretation must start by observing that within this descriptive
trend fluctuations correspond, at least as to direction and timing, to
expectation from our model. By increasing their investments in the
depression of 1921 and in 1922, by decreasing them in the course of 1923,
increasing them again in 1924, keeping them substantially steady during
1925 and 1926, increasing them in 1927, and mildly decreasing them in
1928 and 1929, banks substantially conformed to the prewar pattern of
behavior: there was fairly satisfactory inverse covariation between rates
of change in investments and in loans which, particularly visible after
1 "All Banks" increased their investments from about 5.5 billions in 1914 to about 11.9
in 1919 (see 15th annual report of the Federal Reserve Board, p. 111). The national
banks up to 1917 confined their holdings of governments substantially to what was neces-
sary to cover their note circulation and government deposits. Those in New York City
took the great stride in that year, those outside New York took it in 1917 and 1918 and
reached the maximum at the beginning of 1919.
2 New York banks also liquidated other than government securities. Outside banks
did so only to a quite insignificant extent and only during a few months.
3 And in fact a very substantial increase in loans occurred, as has been noticed in Sec.
E. As has been stated there and will be repeated again — it can never be repeated
enough — this is very material to any rational appraisal of that "deflation." In 1921 the
position of banks was further strengthened by a decrease of currency in circulation which
amounted to 940 millions, by an increase in treasury currency of nearly 220 millions and by
the gold influx (749 millions).
4 Outside banks had, however, only reached about two-thirds of the peak figure of
their holdings of governments by the end of 1922.
6 The figures of June 30, 1928, after which they declined for about 2 years, were 17.8
billions for All Banks. Investments of All Member Banks were about 6 billions in 1921
(June 80) and nearly 10.76 billions in 1928; 10.05 in 1929.
864 BUSINESS CYCLES
the elimination of that descriptive trend, indicates that the former must
have retained something of their old role. We will add the complemen-
tary fact that members (in our sense) also conformed to the prewar
pattern in that they continued to look primarily after their customers'
credit and to make it, even at temporary sacrifices, the pivot of their
business.1 No grasp of banking developments in the twenties is possible
unless this be clearly realized.
But that descriptive trend in investments is entirely due to the three
spurts in 1922, 1924, and 1927, which are obviously associated with
the three major buying campaigns of the reserve system.2 It should
be noticed, however, that this relation is complicated by a number of
other factors, and also that only in the second case did the action of the
Federal Reserve Bank of New York precede the purchasing operations of
member banks. In the two other cases the latter moved first, which
should be sufficient to convince us that these spurts contained an element
which was independent of Federal reserve action: this action supplied
an additional impetus and additional liquid means, but was not the sole
causa efficiens. This is also corroborated by the fact that the net result
of the reserve system's major dealings in governments, from January
1922 to October 1929, was minus 65 millions.3 Nevertheless, it was one
important factor4 — the second or, if we count the war effect, the third —
in shaping members' investments.
The third or fourth factor is the most interesting one. It consists
of the conditions of the banks' customer business adverted to above,
III, b. When banks in and after 1922 saw this business floating away
from them on the tide of abundant money, they did not, strictly speaking,
acquire a new motive to look to investments for a permanent — i.e., more
1 Member banks (in the official sense) are, since October 1928, required to give informa-
tion which enables us to distinguish customers' from open-market loans. This information,
published in the Bulletin and the Annual Reports of the Federal Reserve Board is very
revealing and has been drawn upon in what follows in our text. The point which is
material at the moment is that banks did not turn to temporary investment in the open
market at the expense of their loans to customers when this was more remunerative, as it
was in 1929. This implies that selling and purchasing bonds must also have been, for them,
secondary or subsidiary to customers' loans.
2 But cp. the discussion of the modus operandi of open-market operations which follows
below, sub V.
3 There is always a danger of overestimating the causal role of Federal reserve purchases,
since the system would naturally buy in a situation in which member banks, finding their
funds idling in their hands, would increase their investments independently of any action
by the system. This will presently become still clearer.
4 The present writer entertains a profound respect for any opinion held by that great
economist, A. A. Young, and records dissent with reluctance, but he cannot understand
how Young can have held (op. cit., p. 59) that the open-market purchases of the Federal
Reserve Bank of New York "appear to have no distinguishable effects" on the amount of
government securities held by New York national banks.
1919-1929 865
than temporary — source of employment for their deposit manufacturing
facilities, for as we have seen, this situation was not absent from the
prewar picture. But the old motive gathered additional weight. And
so they increasingly took to holding and handling bonds, outside banks
in particular other than government bonds, which they acquired steadily
even in 1923. The holding provided a return which, though not very
attractive and falling, was supplemented by capital gains and gains
from the handling, especially the placing. So, while still regulating
their course in the way described above, they systematically bought more
when buying than they sold when selling and this is the fundamental
explanation of the descriptive trend we observe. But it also bears on
fluctuations and on the relation of this series to others, for inasmuch as
the method of financing industrial and commercial requirements by the
issue of new or by the sale of old bonds was increasingly resorted to on
the initiative of industrial and commercial concerns, i.e., inasmuch as
it increasingly took the place of direct borrowing from banks, a com-
ponent was inserted that would be positively and not negatively associated
with cyclical phases. We have observed this phenomenon in our survey
of prewar times but it became much more important during the twenties.
We shall expect it to assert itself especially in the holdings of other than
government bonds. So it does — for outside banks it seems to correlate
with business, for New York banks, with speculative activity,1 as it
should. And this explains much of what is unsatisfactory in the negative
association referred to above.
But this negative association was only weakened and not destroyed
by that component. So far as it was not, we shall expect some covaria-
tion between investments and time deposits, since investments due to the
initiative of banks are likely to create idle deposits and idle deposits are —
in spite of the qualifications following from our discussion above — more
likely to be placed on time account than are others. Time deposits
increased at a much greater rate, which accords with our theory that
they do not primarily consist of idle funds, but represent a new way of
handling certain classes of circulating funds, as does the fact that the
logarithmic ratio between investments and time deposits almost steadily
declined. But still there is more than a suggestion of covariation, as
especially in New York figures for 1921-1922 and for 1924. This is what
remains of the grossly erroneous theory that banks invest the time
deposits with which they are "entrusted" by savers. We know that
1 It has been observed by A. A. Young (op. cit., p. 58) that the security holdings of
New York national banks display a variability and quickness of response greater than
those of outside national banks. " The way in which that [the New York] market serves
as a distributing center, holding securities temporarily until they can be absorbed else-
where, is reflected in the character of some of the fluctuations/' This also applies to 1927
to 1929.
866 BUSINESS CYCLES
there is at best fractional truth in the implied relation between time
deposits and savings and that a bank as a matter of principle1 can no
more be said to lend or invest a time than a demand deposit, though to
the banker, provided he adheres to the old phraseology of his trade,2 it
may well look as if he did. Moreover, governments were among the
most liquid of assets and the argument that it would be bad banking to
buy them — or any bonds — except so far as they are balanced by slow
liabilities, sounds strange indeed when it comes from bankers who felt
no compunction when entangling themselves in mortgages.
A number of additional points would have to be dealt with in order
to complete the picture. We confine ourselves to a few of the most
important ones. First, banks improved their balance sheets by increas-
ing their capital. This conversion of deposits into bank stock must
always be kept in mind in any study of the behavior of deposits and in
discussions about the presence or absence of "inflation." It will, how-
ever, suffice to notice the increase which occurred between 1917 and
19213 and which parallels other efforts to normalize their status in
1919 to 1921, in particular the partial liquidation of investments. Sec-
ond, the Federal Reserve Act was bound to change the relation between
pure member banks (our sense) and member banks that also filled
central bank functions and, as far as the former are roughly represented
by outside and the latter by New York banks, to affect the relations
between these two groups. Nevertheless, these relations persisted
rather more than we might have expected: the New York correspondent
retained to a considerable extent his role in the operations of the banks
outside.4 But the transfer of reserve deposits to the Federal reserve
banks affected interbank deposits. The prewar relation between deposits
in New York banks and money in all banks is no longer observable.
Something is left of the prewar relations between New York and outside
net demand deposits and between New York and outside loans and
1 The case of loans on account of others will be mentioned later on.
2 The same excuse cannot be pleaded for theorists who in all other respects have ab-
sorbed the theory of "credit creation." To retain that vestige of old doctrine is in fact
quite illogical.
3 Capital and surplus of national banks stood at 1,845 millions in 1917 and that of
other banks, excluding savings banks, at 1,953 millions. In 1921 capital and surplus of
national banks had risen to 2,300 millions, capital and surplus of other banks to 2,720
millions. But increase continued and in 1928-1929 helped to enhance the banks' lending
power: capital, surplus, and undivided profits of member banks increased by about 2
billions from 1922 to 1929, which item plus total deposits, of course, about equals the
increase in loans plus investments plus reserves.
4 This is shown by the fact, among other things, that rediscounting with correspondents
(New York and others) went on, though on a modest scale, throughout the period. See on
this and cognate points S. E. Harris, Twenty Years of Federal Reserve Policy, vol. II,
App. C, especially the chart on p. 773.
1919-1929 867
discounts. Other points have been noticed already. Third, though the
new organ of central banking and its policy will be discussed later, its
influence on the behavior of banks must be glanced at now.
It was to be expected that the presence of so accessible a source of
additional "funds" would have saved member banks much of their
trouble about liquidity and, hence, profoundly changed their policy.
In the argument of mechanistic theories about the regulation of the
"supply" of money, the assumption that it did forms, in fact, an essential
link. And the growth of real estate loans and other symptoms actually
point in that direction.1 But, no matter whether by compulsion from
the federal reserve authorities, by tradition, or by perception of the
long-run advantages of the observation of "sound" principles, the banks
yielded to such temptation as there was much less than one might
think. To be sure, their behavior varied greatly, especially as between
first- and second-rate institutions.2 The recurrence in the Federal
Reserve Board's annual reports of passages suggestive of difficulties
with banks which tended to lean on the snaffle in order to increase their
working resources or at least to profit from differences between open-
market and rediscount rates, is no doubt significant. On the whole,
however, there cannot have been very much of this. The modal and
still more the high-class bank did not relish being indebted to its reserve
correspondent and was usually anxious to reduce its debt.3 Looking,
as has been remarked above, primarily after its customers' business, it
more majorum so regulated open-market commitments and investments
as to be able to finance it without falling back upon rediscounting save in
exceptional circumstances. This can be seen clearly only since October
1928, when the necessary information first becomes available: in the
last year of the period member banks in the aggregate actually increased
their customers' loans, while at the same time paying back that incre-
ment of debt which they had incurred at the reserve banks in the first
half of 1928, and they achieved this, until October 1929, in the face of the
1 In this connection it should be mentioned, however, that member banks' reserve
balances increased, January 1922 to January 1927, by nearly 560 millions (though " Federal
reserve credit outstanding" declined by 140 and currency in circulation increased by 375
millions) largely as a result of gold imports. Reserve balances were (monthly average)
1.742 billions in 1919, 1.655 billions in 1921, and 2.375 billions in 1929. Some authors
regard this increase of about 700 millions as the pivotal figure in the monetary processes
of the period, in particular as proof positive of the presence of "inflation."
2 Since the largest banks happened to be located in New York, that difference may,
roughly and in the aggregate, be said to correspond to a difference in behavior of New York
and outside banks. See L. Currie, op. cit.> pp. 91 et seq.
3 In the case of New York banks, increase in indebtedness is, in fact, regularly followed
by a decrease in net demand deposits, as Dr. Currie has pointed out (op. cit., p. 93). But
it must not be forgotten that there are also other reasons, originating in the cyclical rhythm,
which would tend to produce that result, irrespective of reluctance to being in debt.
868 BUSINESS CYCLES
reserve banks' sales of governments, by reducing their loans to brokers
and their holdings of acceptances and commercial paper, though the
simultaneous influx of gold greatly helped. But the behavior of redis-
counts (see below) suggests that the inference may be generalized so as
to cover the whole period.1 It follows that banks, assisted no doubt by
exceptionally favorable circumstances, were never "loaned up" except
in the irrelevant sense of that phrase which includes secondary reserves.
In this irrelevant sense they may have been much of the time, though
even in the first quarter of 1929 there were small excess reserve's. They
never were loaned up in the relevant sense, i.e., in the sense that without
open-market operations they would have been embarrassed by a customer
asking for a loan. We therefore conclude that member-bank practice
and hence the meaning of member-bank figures after all differed less
from prewar patterns, and that the figures are more amenable to interpre-
tation in terms of the processes embodied in our model than we might
have expected. This must be borne in mind when we try to appraise
the role in the processes of the twenties of the increase of nearly 700 mil-
lions (1922 to 1929) in the reserves of all member banks, or of the increase
of about 11.5 billions in their total loans plus investments.
d. Discussion of the struggle of German banks with peculiarly
German postwar conditions would present many points of interest but
would not add materially to our understanding of the postwar cyclical
process. The most important feature, foreign credits, has been fully
dealt with earlier in this chapter. These, of course, powerfully influenced
the behavior of "deposits and circulation" (Chart XL), which also
reflects gradual return, after inflation, to normal habits of holding bal-
ances and the Juglar phases. "Creditors" in leading banks (only very
roughly comparable to demand plus time deposits in the United States),
which had been as low as 3.8 billion reichsmarks in 1924, rose to about
12 billions in 1929 but there is, under the circumstances, nothing remark-
able in this increase and, in particular, no indication about the rate of
saving; if anything it indicates disproportionate expansion of monetary
purchasing power. The Debitoren in current account, with which we
include advances and bills held (this is very roughly equivalent to all
other loans), increased steadily, though with the expected variations in
rate, from a little over 3.3 billions in 1924 to nearly 9.4 in 1929 and show
that orthodox bank credit did not in Germany lose nearly so much
ground as it did in this country. The same can be inferred from the
1 The above is perfectly compatible with some covariation between rediscounts and
outside loans and discounts which, but for open-market operations, would be still more
clearly marked. They rose together in the postwar boom, fell together to the middle of
1922, rose together in the revival of 1922-1923 and then again in 1925 and 1928. Only
1924 shows marked, 1927 a less marked, divergence that is easily accounted for.
1919-1929
869
series, which continued to be of fundamental importance, of Bills Drawn,
although their share in the total of credit outstanding rapidly declined.1
Money in circulation also displays a rising tendency to 1929 but surpassed
the 1913 figure only insignificantly in that year. Other aspects will be
noticed later
1919 1320 1921 1922 1923 1924 1925 1926 1927 1928 1929 I9SO 1931 1932 1933 1934 1935 1936
CHART LI. — London clearing banks' figures (see Appendix, p. 1073).
All that it is essential for our purpose to observe about the English
case can be seen at a glance in Chart LI.
The striking difference between this picture and the American and
German ones, due to the " classic " lines of the former, needs no comment
beyond a reminder that those lines were the result not only of the mone-
tary and fiscal policy pursued but also of properties of the English
economic process which were completely independent of them. Treasury
1 The amount of bills held by banks increased almost steadily whether the total in-
creased or decreased. This shows, if proof be wanted, how very liquid they were all the
time.
870 BUSINESS CYCLES
Bills Discounted have been inserted in order to avoid missing the fact
that public financing was throughout the central factor in the money
market around which everything else turned. Advances display a most
orthodox inverse covariation with investments, which during the period
never acquired the significance of their American counterpart. Be it
repeated that current accounts1 kept up in 1920 and the first half of 1921,
while a clearings figure which is supposed to parallel American outside
debits declined to a new and very stable level. Moreover, they display
but the most gentle suggestion of cyclical fluctuations: Juglar phases are
just recognizable in the fact that they slowly fell to 1925, then ceased to
fall, and in 1927 and 1928 slightly rose. It is, however, interesting to
note that, as indicated by the curve at the bottom of the chart, that slow
decline is not paralleled by clearings, the indicators of monetary business
volume, and hence cannot have greatly affected business processes.
IV. The partial description of the monetary aspects of the invest-
ment process in the twenties, which is implicitly contained in the analysis
just presented (III), must now, in some points at least, be made explicit
and also supplemented by other data from the financial sector, mainly
the stock market.
a. For this purpose it is first of all necessary to insert into the picture
the phenomenon which so much increased in importance during the first
postwar decade — everywhere, but in particular in this country — Tempo-
rary Investment.2 We have seen that industrial concerns which happen
to hold owned funds3 in excess of what they require at the moment — for
instance, undivided profits or proceeds from the issue of bonds or stock —
insurance companies, public bodies, wealthy individuals in the act of
shifting their investments, often find it convenient to acquire quick
assets of the type that banks acquire for secondary reserve purposes, or
even assets which a bank could and would not buy for this or any other
purpose, and so to take, for a time or regularly, a hand in the game of the
open market. Instead of buying such things as acceptances, treasury
bills, or even bonds, they may, of course, also lend in the term's most
narrow sense, e.g., to bill brokers or stockbrokers. The way in which
this special case of temporary investment affects the situation of banks
does not differ from the way in which it would be affected by the pur-
1 As has been mentioned before, the English distinction between current and deposit
account seems to the writer to be somewhat more significant than the American distinction
between demand and time deposits. This is why for the curve at the bottom of the chart
clearings are divided by current accounts only.
a The term is to be understood in the technical sense given to it in Chaps. Ill and XI.
8 Though confining ourselves to owned funds, we should remember that the phe-
nomenon itself is not necessarily confined to owned funds in our sense. It should also
be remembered that there is no contradiction in saying that, say, an insurance company
practices temporary investment permanently.
1919-1929 871
chase of bonds; in fact, we may usefully call the transaction "buying a
loan." If the loan bought has already been in existence and "owned"
by the bank which holds the deposit to be thus invested, then this bank's
deposits and loans are reduced by equal amounts so that the pressure, if
any, on its cash or its reserve deposit is, ceteris paribus, relieved and its
lending power increased.1 The banking system's aggregate lending
power is also increased and we may say that existing deposits now go
further or that their "velocity" rises. If the loan did not exist before
but has been newly created by the transaction, the recipient being still
assumed to bank with the same bank as the lender, the bank's deposits
and loans are not affected and its lending power is not increased.2 If
either the purchase of the preexisting loan or the emergence of a new
one leads to the transfer of the corresponding amount of the investor's
balance to another bank, as in the United States it mostly would, we
get an additional feature which for the investor's bank is in the nature
of a shock. There are several obvious ways to absorb the latter, such
as using elbow room the bank may have been sensible enough to pre-
serve, or borrowing at its reserve bank or from the receiving bank.
Moreover, in the case of actual transfer of reserve balances, the status
of the receiving bank will be improved and its lending power increased,
normally by the same amount by which the paying bank's lending power
is curtailed. But no generalization about this can have any realistic
virtue. Net contraction of aggregate loans in the banking system is
perfectly possible, especially if the transfer happens to be from a bank
in a strained position to a bank in a comfortable one, though even in this
case the ultimate effect for the economy as a whole will be much akin
to that of an expansion of bank loans.
So far we have dealt with temporary investments or, to take the
special case, loans by nonbanks within a closed economy. The logic is
no different if newly imported funds are used for the purpose, but there
is a practical difference in the complex of effects if immigration of those
funds entails an addition to member bank reserves. Let us imagine, in
order to illustrate this point, that a foreigner (in the twenties) dropped
from the sky with the equivalent of a million dollars in gold which he had
salvaged from oppressive taxation or anticipated political trouble in
1 Cetera, however, need not be paria. But assuming that the bank is now able to
make additional loans, it is worth while observing that, in this as in some other cases,
decrease in deposits will be accompanied by increasing ease in the money market and a
tendency for money rates to fall. This otherwise trivial observation acquires some impor-
tance in the presence of theories that associate rise and fall in deposits with falling and
rising tendencies in interest rates.
2 It might be said, however, that it is potentially increased, if the recipient of that loan
would otherwise have applied to the bank, a very realistic case : his impending application
being warded off, the bank can now view other applications more favorably.
872 BUSINESS CYCLES
some other country, and that he therewith acquired a balance in some
member bank. This, of course, would swell that bank's reserve with its
Federal reserve bank, and whatever the depositor does with the balance
he acquires — we may well imagine that he will invest temporarily until
he has recovered from his fall — the effects of the incident increase in bank
reserves will overshadow any effects his action may have.
The motives of temporary investors are clear. There are, however,
also motives for them to use their banks as intermediaries — motives of
technique and others: the bank may, for example, be a useful screen.
Banks complied in the first place because they had to. Temporary
investors are, of course, competitors and their very presence was a
symptom of the weakness of the banks' control over the financial struc-
ture. But in the second place, a bank stood to gain more than the
commission by having the placement of such funds left to it. In par-
ticular, it could then use them in order to relieve its individual position
and also in order to finance purposes in which it may have been directly
or indirectly interested without caring, in times of active criticism, to
sponsor them too publicly suo nomine. So cooperation was indicated.
It did not alter the modus operandi of these loans, but it entailed the
statistical consequence that they were now reported as Loans on Account
of Others. Since they were mostly directed toward the stock exchange,
they showed up as part of Brokers' Loans,1 while banks gave another
part on their own account, as a matter of their business routine.
Another remark will complete the theory of this special case of a
special case of Temporary Investment and, incidentally, our discussion
in Chap. XIII, Sec. D. In practice, a stockbroker must not only have
assets of his own but also keep a good bank balance. We may, however,
assume that he will borrow part of this balance against his other assets
so that his mere readiness to do business will produce some brokers' loans.
If speculators now do what in Chap. XIII we have called "buying them-
selves in," i.e., transfer to brokers certain sums to start their margin
accounts — they need not do this, however, but can deposit securities
instead — these sums, though they may also be borrowed, will diminish
the total of brokers' loans. The aggregate of the remainder will not
change when brokers thereupon get busy with executing orders to buy
or to sell.2 For though individual brokers will have to borrow in order
1 That part includes, for any individual bank, also what it thus lent on account of other
(out-of-town) banks. The behavior, of this element was in between the behavior of
brokers' loans on account of nonbank customers and of brokers' loans on the lending
banks' own account. Our argument primarily refers to the two latter categories.
* This point is overlooked surprisingly often. The writer is not very familiar with
the literature of stock exchange operations, but as far as he knows, credit for stressing
it is due to Mr. W. J. Eiteman, Economics of Brokers' Loans, American Economic Review
for March 1932.
1919-1929 873
to replenish their own deposits as soon as they meet with an adverse
balance on the day's transactions, other brokers must, ipso facto, acquire
a favorable balance which up to the amount of their indebtedness will
reduce brokers' loans by as much as they have been increased by the
borrowing of the buying brokers, unless it be withdrawn by customers.
It is true that this will be done very promptly, especially if the sellers
are not speculators but either the issuers of the securities they sold or
outsiders without margin accounts. Nevertheless, it is neither buying
nor holding as such that brokers have corporatively to finance, but
withdrawals or, as we may also put it, conversion of balances with brokers
into balances with banks. Hence, although brokers' debts are technically
incurred in order to service customers' debts, aggregate brokers' loans
can in general only increase through conversion of customers' claims.
This conversion might create a very awkward situation. The con-
stant threat of it is, in fact, one of the brakes that are essential to the
normal working of the money market mechanism. But if a large amount
of owned balances stands ready to "buy" the claims against brokers,1
new balances need not, to that extent, be created, and speculators'
original deposits with brokers plus profits or minus loss can be piloted
into the channels of consumers' and producers' expenditure without
for the time being causing any disturbance — to the banks or anyone else —
except so much as is implied in hitherto idle deposits becoming active.
We cannot follow up the interesting implications of this argument, but
it should be observed that it almost completely disposes of the theory —
held not only by politicians and popular writers but apparently also
by some economists and the Federal Reserve Board — that brokers' loans
during the twenties absorbed funds that "legitimate business" was
supposed to need sorely. On the contrary, brokers' loans were a means
of, as it were, coining speculators' gains and injecting them into the
stream of economic life. They might, therefore, be called "inflationary"
in an obvious sense of that term. And an argument against them can be
forged from that.2 But this does not save the other.
It is in this light, as a concomitant or vehicle of realized speculators'
gains or of stock issues, that we must look upon brokers' loans. They
increased strongly in 1921-1922, 1924-1925, and 1927-1929. But what
1 For that is what lending to brokers comes to. From the standpoint of banks, how-
ever these loans are newly created. It should be borne in mind that the argument applies
equally, with little modification, to the case in which the sellers are the issuers of the stock
as to the case of transactions in "old" stocks.
2 See Chap. XIII, Sec. D. It is, in fact, not farfetched to surmise that this is what the
Federal Reserve Board meant when it tried to fight speculation. It may also have resented
having its wires crossed by lenders beyond its control and having to listen to bank execu-
tives declaring in reply to admonitions that these lenders were as independent of them
(the executives) as the man in the moon, which, of course, was not quite so.
874 BUSINESS CYCLES
banks contributed on their own account was not only comparatively
steady — around 1 billion even during the last three years1 — but actually
tended to decline through the greater part of 1928 and, after a strong
rise in the last quarter of that year, again during the first half of 1929.
The spectacular ascent in 1928 and 1929 to almost 7 billions was entirely
due to loans on account of others, which from the beginning of 1928
to the maximum in 1929 rose from about 1 to nearly 4 billions.2
b. We will now inspect Chart LII.
Bearing in mind the argument just presented, we shall not interpret
the covariation between brokers' loans and prices of industrial stocks,
obvious and understandable as it is, in the usual way, i.e.* in the sense
that the abundance of funds available for stock exchange operations
was the initiating force of the boom: in part, especially in the case of
foreign funds, it was, if anything, the consequence of it. The relation,
however, is more complicated than that. The initiating force was, as
always, industrial success. But the abundance of money was a condi-
tion for speculators' finding it so easy to convert gains into balances and
for concerns' being able to sell such quantities of new securities at high
prices. Relending part of the proceeds, which in the last years were
much above immediate requirements, in turn increased brokers' loans.
This practice further helped to relieve banks of the pressure which
conversion would otherwise have put upon them, and to paralyze for a
time the brakes of the engine. Conversion thus facilitated further
conversion instead of impeding it, as it would normally do, and abundance
of "funds" created additional abundance.
We may note that these "funds" would, on their way to and from,
not necessarily but frequently pass through time accounts — thus effecting
the utmost economy in the use of demand deposits — and that converted
funds — or funds created by conversion — which were neither redirected
to brokers nor immediately applied to financing consumption or real
investment would regularly be placed on time account. This is the only
nontrivial relation between time deposits and stock prices, i.e., the only
relation which means more than that both happened to increase under
the influence of environmental conditions common to them. And it
has nothing to do with the investment of savings; hence, cannot verify
any theories about it.
The above analysis may be translated into terms of money rates.
The theorists who establish, via the rates of interest, a simple causal
1 A sharp rise, followed by a still sharper fall within a few weeks, occurred, however,
in the fall of 1929 (see next footnote). The figures given are those of New York City
member banks, also for the loans on account of others.
2 The remainder consisted of loans on account of out-of-town banks. The 7 billions
refer, it should be remembered, to reporting member banks. The grand total was, for
September 1929, about 8.5 billions.
1919-1929
875
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART LII.— United States (see Appendix, p. 1073).
876 BUSINESS CYCLES
connection between abundant money and booming stock markets cannot
glean much comfort from the course of events in the twenties. It is true
that the mechanism described kept money rates lower than they would
otherwise have been. But the struggle for conversion, though facilitated
to the utmost, all the same asserted itself in a relatively high call rate,
which did not, however, spell tight money all round. As to the first
point, we observe that only to the end of the first quarter of 1925 call
rate kept anything like its regular relation to other open-market rates,
especially the commercial paper rate, i.e., the relation which would
follow from its place in the structure of the banking business. Till
then it was indeed (almost) consistently lower, reaching its (monthly)
trough for August 1924 when it was 2 per cent, while commercial paper
rate still stood at 3.25. But afterward it was (almost) consistently and
increasingly higher, the difference reaching a maximum (3.71) in March
1929. The presence of a stimulus to speculation from low rates could
hence — even if we waive theoretical objections — be averred only for the
second part of 1924. As to the second point, open-market rates, of
course, fluctuated together. And, as we have seen (Chart XLIV),
interest on customers' lines of credit increased in 1928-1929 by, on a
rough average, 1 per cent. But, as we have also seen, no restriction
took place in that quarter.1 Bond yield reacted but moderately (see
chart). And money raised by the issue of stock was never so plentiful.
The "strain" of 1928 and 1929 was substantially confined to the stock
exchange. It came to a head, as it was bound to, through the retreat
rather than the exhaustion of the funds which financed brokers' loans and
which through them, in the way indicated (see above, sub a), served the
purposes of speculators and issuers, while at the same time they were
by their owners (other issuers among them) earmarked and counted
upon for other expenditure. In the end, fright may have been what
turned retreat into rout. And that fright may have been partly moti-
vated by the perception of absurdly high stock prices and may thus link up
with the fundamental fact that the economic process was heading toward
1 It might be replied that an increase in the interest charge is restriction. What
the statement in the text means is, first, that banks did not restrict loans to business. As
has been pointed out, they increased them and were enabled to do so by various favorable
circumstances and by their own action, i.e., by reducing their open-market commitments.
Second, we hold that in booming conditions a 1 per cent increase in interest charges does
not induce significant restriction of business operation. Perhaps, in strict logic, it should;
but it does not. The economic world is no billiard table. What, then, did that increase
mean? Nothing except that banks will increase their charges if circumstances give them
an excuse for doing so. The monetary strain of 1928 and 1929 is pure moonshine. Or
rather, he who speaks of strain betrays that, when speaking of business, he is thinking of
margin accounts. When in 1929 the funds of "others" took fright, the New York reporting
member banks stepped into the breach so as to increase their demand deposits by nearly
1.6 billions. How could that have been done if there had been any strain beforehand?
1919-1929 877
depression. But the situation in the stock market was untenable
independently of all that. Precisely if business had gone on booming, the
nonbank lenders would themselves eventually have needed their funds.
The pyramid of loans on account of others, and stock prices with it,
would have had to come down for this reason alone.1
Some minor points in the picture may be noted. Prices of railroad
stocks increased materially less than the prices of industrials. Yet the
writer, being unable to see anything but gloom in the situation of rail-
roads, finds it difficult to understand why they should have risen as
much as they did. New York debits reflect the transactions of the stock
market well though, owing to obviation and the fact that its relative
importance is bound to increase in consequence of an increase in the total
of transactions, on a progressively reduced scale. Temporary invest-
ment— brokers' loans, in particular — explains the quiet behavior of
New York City bank loans and net deposits.
No great value attaches to the data about issues for the purpose of
securing new capital, embodied in Chart LIII.
We may, however, note a few significant features. There is, first, the
ominous increase in the flotations of securities of investment trusts and
financial and trading companies since 1926, which reflects the doings
of the last three years of the period and teaches much about the specifi-
cally American characteristics of the great crisis. The role of the Land
and Buildings group should next be noted — much of it was of a similar
type and no less indicative of troubles to come. Third, the industries
that carried the third and fourth Juglars are mostly represented, as we
should expect, though some of them very modestly. Public utilities
stand out as the great consumer of capital. Fourth, if we discard the
classes that directly link up with the mania, we see the rise of the fourth
Juglar clearly indicated. If it was not more so, this was due to the fact
that, as stated before, concerns took advantage of the opportunities the
stock markets offered to them 1928-1929, even if their immediate
investment programs did not require any outside financing.2
1 The argument will be developed from another standpoint in V. The reader should
note both the superficial similarity and the fundamental difference between that argument
and the Karsten theory, see Journal of the American Statistical Association for December
1926. Also cf. Persons and Frickey, Money and Security Prices, Review of Economic
Statistics for January 1926. Professor Bresciani-Turroni's important contribution should
again be mentioned: Considerazioni sui Barometri Economics, Oiornale degli Economists
for January, May, and July 1928. Finally, comparison is invited with the analysis of
the boom by Professor Lindley M. Fraser, American Economic Review for June 1982.
2 The evidence of the chart may be usefully compared with the increase in total long-
term debt. That of railroads was 11.9 billions in 1922 and 13.4 in 1930, that of utilities
was respectively 8.4 and 14, industrial long-term debt increased from 6.8 to 10.8 (U. S,
Dept. of Commerce, Long-term Debt in the United States, p. 6).
878
BUSINESS CYCLES
Millions of Dollars
1,200
2,100-
1920
1,500
1925
2^00
1921
1922
1923
1924
1926
1927
1928
0 inn
D RAILWAYS
1,800 -
1,500 -
1200 -
D PUBLIC UTILITIES
• i IRON, STEEL,COAL, COPPER
H EQUIPMENT MFGERS.
8 MOTORS AND ACCESSORIES
: I LAND, BLDGS. ETC
li RUBBER
900-
• INVEST. TRUSTS,TRADING, ETC
600-
300
0L
r
1
1 1
;
Ui
ill [
ill I
1929 1930 1931 1932 1933
CHART LIII. — New capital issues; New York Stock Exchange (see Appendix, p. 1074).
1919-1929
879
An investigation made in order to get at figures more relevant to the
processes of real investment,1 reveals the fact that in 1929, out of the
total of public issues by domestic corporations, which amounted to
nearly 9.4 billions, at the most % billions can have been (directly) for
the purpose of providing real, or " nonfinancial " or "productive," capital,
1920 1925 1930
CHART LI V.— United States (see Appendix, p. 1074.)
1935
and even these £ billions include issues for providing working capital,
the greater part of the issues of installment finance and real estate loan
companies and issues the purposes of which were so ambiguously stated
as to make inclusion or exclusion a mere tossup. One of the results,
however, was to strengthen confidence in the reliability of the series
1 It soon proved too laborious to be extended beyond the year first tackled, 1929, but
has been published recently. See George A. Eddy, Security Issues and Real Investment
in 19129, Review of Economic Statistics for May 1937. Perusal of the appendix of that study
will introduce the reader to the difficulties of the problem. Issues for the purpose of
repayment of bank loans have been considered as "productive" if those loans themselves
were productive and not older than three years.
880 BUSINESS CYCLES
of "productive" issues published by Moody's Investment Survey.1
This has, accordingly, been used in Chart LIV.
The order of magnitude of, and the cyclical fluctuations in, these
"productive" issues are highly interesting and in themselves sufficient
to dispel many errors about the investment process of the twenties.
The annual figures rise from 864 millions in 1921 to 1,941 millions in
1924, the maximum for the period under survey (though it was nearly
reached again in 1930). Then they fall, the figure for 1928 being below
that for 1924 by 446 millions. The influence of the speculative excesses
of 1929 shows, but the figure for that year is still only 1,787 millions.
This is much more like what we should have expected from our report
on the industrial processes of the time. The perfect independence
from the course of interest rates, bond yields for instance, should be
particularly noticed. It would of course be hazardous to aver that these
issues are a perfectly reliable indicator of investment in general — in
fact, it is certain that corporate issues will in general differ in timing from
other methods of financing real investment and also from the total of
actual expenditure. There is anything but satisfactory covariation
between them and the other series plotted on the chart.
c. Normality (conformity to our scheme) is what strikes us at first
glance when we inspect the German chart (Chart LV) .
We must make a partial exception for Foundations and Failures,
which increase and decrease in 1928—1929 in a manner not in accord with
other indices or with expectations. But call rate, though, of course, not
absolutely lower than in the United States, normalized its position rela-
tively to other rates in 1925 and later on the whole kept this place except
during the first 7 months of 1927 and then again in 1929. The boom
of 1926-1927 as reflected in stock prices and stock exchange transactions
was part of the processes of the Juglar prosperity though, within it, it
came later than is usual. After that we observe the decline, also in
dividends and issues, which in a Juglar recession on a Kondratieff depres-
sion is the normal thing to expect. And in fact, although there was
much trouble ahead, within as well as without the sphere of banking, the
stock market as such never became a center of difficulties in the subse-
quent great depression. Yet neither ease in the open market nor loans
on account of others were absent. The influx of foreign credits created
conditions not dissimilar to those prevailing in the United States. The
ease from 1925 to 1927 was closely linked up with such loans, the pro-
1 Vol. 25, No. 86, October 1933, p. 1671. This series excludes additions to working
capital and thus, including substantially the items plant and equipment only, really
serves our purpose better. With it, Moody's Survey combines another, which includes
municipal and farm loan issues, which are also considered as productive. The total is
what has been plotted, but the figures given in the text refer to (domestic) corporate
issues only.
1919-1929
881
ceeds of the issue of industrial bonds and stocks — see chart ; total domes-
tic issues amounted to about 3 billion reichsmarks in the first 9 months
of 1926 — being partly used for the repayment of short debts to banks or
for loans to the market. Correspondingly, the stock exchange loans of
1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART LV. — Germany (see Appendix, p. 1075).
leading banks (Reports and Lombards), after increasing rapidly during
1926, reached a maximum of 1 billion at the beginning of 1927. But the
Reichsbank which, throughout Mr. Schacht's tenure of office, ener-
getically exerted what regulating influence it had, really meant business.
And the severe lesson administered to speculation in 1927 proved suffi-
cient to stop the upward cumulative process and to prevent its recrudes-
882 BUSINESS CYCLES
cence for the rest of the period, hence a crash of the American type1 at
the end of it.
In the London stock market conditions of boom prevailed, almost
without any setback, from the middle of 1925 to the middle of 1928,
after which there were upper-turning-point hesitations for about one
year. See Chart LVI.
This boom, normal within the cyclical schema and also reflected in
"Town" clearings, deserves some emphasis because of the divestment
by banks which accompanied it throughout (see Chart LI) and because
of the fact that it followed upon the heels of the gold standard act. From
our standpoint there is nothing to wonder at in this. As a matter of
mechanism, however, it should also be observed that divestment corre-
sponded in timing to the strong increase in the London clearing banks'
loans at short notice to the stock exchange. Moreover, although there
are a number of shortest-run exceptions, the inverse relation of the
movements in stock prices and in call rate is almost perfect. The short-
money index2 fell from its maximum of 163 per cent of the 1913 average
1 This is all that the limitations of our purpose and the writer's wish to deal more
adequately with American developments will allow him to say about the financial aspects
of the German investment process and the policy of the Reichsbank during those critical
years. In dismissing the subject, he wishes to add three remarks. First, in view of the
host of abnormalities in the German situation, the normal complexion of the sector under
discussion must, in fact, be largely attributed to the bold handling of the Reichsbank,
though control was made easier by the factors which darkened the industrial outlook after
1927. Second, the policy of the Reichsbank no doubt presupposed not being afraid of
timely use of bank rate. But, partly because bank rate is never the decisive element in
any situation, partly because owing to the foreign credits it could not have had full effect
in this situation, such use as was made of it was only of secondary importance. The latter
difficulty, expressing itself in the unwieldy mass of loans on account of others, was never
really overcome, though steadfastly fought. It is highly instructive for anyone, and
should be even more so to believers in the key position of interest rates, to recall that in
June 1927 the first reaction of the stock market to the decision of the Reichsbank to raise
its rate from 5 to 6 per cent was a rise in stock prices, it being argued that this would draw
foreign funds: "stocks firm on dear money" was the slogan of the hour. Hence, third,
it was not by raising bank rate but by forcing banks to withdraw what corresponded to
brokers' loans on their own account which did the job. It is true that the measure took
effect so promptly because in Germany the loans on account of others were not, as a year
later they were in the United States, almost exclusively directed toward the stock exchange.
But it would have been effective even in this case. For however little banks may on their
own account contribute, that little is of cardinal importance and its withdrawal would in
general be quite sufficient for any desired degree of damping. This is not without some
bearing on the responsibility of banks for booms not directly financed by them : the officers
of a regular army may be entirely innocent of the excesses committed by irregulars over
whom they really may have no control; but if they use however small a detachment of
their regulars in order to provide rallying points for the irregulars, they can no longer
decline responsibility.
2 International Abstract of Economic Statistics, International Conference of Economic
Services, 1934, pp. 105-107.
1919-1929
883
to 48 per cent of it in July 1922, then rose and kept well above that
average during 1926. In 1927 it fell below it, however, and it was only in
1929 that, under the influence of New York events, it rose to anything
like abnormal heights. The hesitations of the curve of stock prices
set in while it was still, falling and bank rate no higher than it had been
M
I
/i
OCK
A\/^
JON CALL RATE W
V/
EY A
O ST)CK
\N
Or
SHORT NOTICE
XCHX NGE
A
V
\r
1919 1920 1921 1922 1323 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
CHART LVI.~ England (see Appendix, p. 1075).
through the boom. In what is an instructive contrast to the American
case the subsequent crash, or as much of it as was not a repercussion
from New York, was more a crash of unsound concerns (Hatry) than a
breakdown of a structure.
To the policy of the Bank we shall return. But a graphic survey
of flotations may be added here. The strict regulation of issues, which
884
BUSINESS CYCLES
Millions of «£ n
20
70-
60-
50-
30-
20-
1897
1903
1898
1899
1900
1904
1905
1906
1901
1902
1907
1908
Kf)
40[~
30
20
10
[] RAILWAYS
1 MINES
0 PUBLIC UTILITIES
r
| SHIPPING
A
i ESTATE AND LAND
i] i MOTORS
^yjlJj i RUBBER
1913 i OIL
:;* - p
In i
911
i'.
1
Li
912
1909 1910
10 -=•
1925 1926 1927 1928 1929 1930
CHART LVII. — English new capital issues (see Appendix, p. 1076).
1919-1929 885
for part of the period amounted to prohibition, must of course be kept in
mind. In order to show up the postwar features of the picture by con-
trast, the survey has been extended over the whole Kondratieff. The
chart speaks for itself (Chart LVII).
The path that leads from the financial sector to real investment is
tortuous and unsafe. Recalling, however, what has been said about the
subject on various occasions earlier in this chapter, and once more
accepting Mr. Colin Clark's guidance, we cannot doubt the broad fact
that investment in the twenties was a much smaller percentage of national
income than in 1907. 1 The expansion of production, which as we have
seen was considerable, in fact took place along with a great relative
decrease in real as well as monetary accumulation. To conclude from
this that the rate of economic growth does not depend upon the rate of
accumulation of "capital," is not more but also not less reasonable than
it would be to conclude that output does not depend upon the degree to
which resources are utilized. We know that change in production func-
tions is more important than either. And one-variable relations do not
work any better in this case than they do in any others. But consider-
ations such as these are an unsafe basis for the conclusion which econo-
mists of the anti-saving schools would evidently love to draw from them.
V. a. That cross between European doctrines about banking and the
exigencies of the American environment which is or was the Federal
Reserve System,2 was from the first endowed with, and in 1917
acquired additional, powers of expansion, little in accord with those
doctrines. Circumstances too familiar to insist upon prevented them
from acting on foreign exchanges, but in all other respects the "economy "
of reserves effected3 by the act and the amendment amounted to a
1 National Income and Outlay, Table 84, p. 185. Mr. Clark's figures are 235 million
pounds for 1924 and 255 for 1929. Overseas investment fell below the prewar level even
absolutely. But his definition of Net Fixed Capital is much wider than ours. The figure
that is relevant for us is more akin to Mr. Clark's Industrial and Commercial Capital
Outlay, which was respectively 81 and 72 millions (Table 88, p. 193).
2 The influence of a prominent banker whose thoughts on the subject were the product
of German tradition — which, of course, means that in part they were also English — is
clearly traceable. There are clauses in the act that are suggestive of the old Reichsbank-
gesetz — another instance of the truth that nobody disregards facts so completely as the
theorizing practitioner, an instance also for the other truth that any enactment is eventually
adapted by the environment.
8 The "economy" was brought about, first, by direct reduction of reserve requirements,
second by the concentration of member bank reserves at the Federal reserve banks. Esti-
mates of the amount of the economy due to direct reduction would vary somewhat accord-
ing to the hypotheses adopted as to the modus operandi of the measure. An estimate that
has been widely accepted, that of the Federal Reserve Bank of Richmond, puts the average
reserve requirements of all banks before the Federal Reserve Act at 21.09 per cent and
after the amendment of 1917, at 9.76. The special reduction for time deposits first to 5,
then to 3 per cent acquires enhanced importance if our view of the nature of time deposits
886 BUSINESS CYCLES
devaluation of the dollar. Like any devaluation, it might have remained
dormant for an indefinite time. But the war made it effective in a way
which should gladden the heart of the quantity theorists: from 1914 to
1920 deposits in All Banks approximately doubled — they increased from
about 18.6 billions to about 37.7 — and the BLS index of wholesale
prices also roughly doubled between 1914 and 1919, which year, owing
to the somewhat speculative character of many of the prices which enter
into any wholesale price index, it is but fair to choose for a test.1 The
expansive power of the system was not, however, exhausted thereby.
As evidence from time series already presented shows, it was perfectly
up to financing at the new level of prices and values and, as has been
pointed out before, able not only to stay as expanded by the war but also
to expand further. Two additional reasons for this should be mentioned
her<j,,
First, the ratio between currency in circulation (outside the Treasury
and All Banks) and total deposits "rose sharply in the two war years,
but then resumed that fairly steady decline which goes back at least to
1893, and resumed it at nearly the same average rate of relative decline
per year as before."2 Evidently that process which we have dubbed
the immigration of legal tender into banks and which was so important
an element of the expansion of bank credit in the past, had not been
completed. Currency outside of the Treasury and All Banks, as far as
the annual figures reported by the Comptroller of the Currency allow
us to judge, increased rapidly to a maximum for 1920 and then fell
be accepted. The conversion of member bank reserves into deposits at the Federal reserve
banks against which the latter were only required to hold 35 per cent of "cash," and replace-
ment of gold certificates by Federal reserve notes covered to 40 per cent in gold amounted
to a further reduction of reserve requirements for the banking system as a whole and
multiplied the potential effect of the reduction of member bank requirements, thus provid-
ing the machinery for "classic" war inflation. The policy temporarily adopted of paying
out gold certificates in preference to notes, of course, mitigated pro tanto the last-mentioned
effect.
1 The reader is presumably by now aware of the fact that the writer does not harbor
any too friendly feelings toward the quantity theory and, in particular, its modern revival.
The fact that it works comparatively well under the circumstances of war inflation pre-
cisely shows that it is worse than useless if made to stand on its own legs, i.e., on the logic
of the equation of exchange as such. But we must give the devil his due and cannot agree
with those economists who try to deny or explain away that fact itself. Exact covariation
of prices and quantity of money it is, of course, unreasonable to expect. And foreign
demand for American commodities does not provide independent explanation but was part
of the very mechanism that the quantity theory assumes to exist.
2 J. W. Angell, Behavior of Money, p. 17. The downward drift, though less pro-
nounced, is also present in the ratio between outside currency and adjusted deposits subject
to check (ibid., p. 19). Attention should be drawn to the table at the foot of p. 17 which
shows that the English ratio, while more than twice as high, also shows the decline and at
approximately the same rate.
1919-1929 887
rapidly to a minimum for 1922, the difference being roughly 1 billion.
The figure for 1923 shows a " normalizing " increase, but 1924 and 1925
show decreases and 1928 and 1929 but small increases: cyclical drains
and reinigrations are very feebly marked. We have a seasonally adjusted
monthly series for currency outside the Treasury and Federal reserve
banks which moves better in the cyclical phases, indicating the rise of
the fourth Juglar, falling in 1927, keeping constant in 1928, falling in the
first half of 1929 and again, after an understandable upward jerk, toward
the end of that year and almost throughout 1930. This is important
to keep in mind, considering the role in the cyclical mechanism attributed
by some theories to cash movements. For us the point that matters is
the accession of power to expand which this development brought to
member banks as well as to reserve banks.
Second, while it is right and proper to stress, as most economists do,
the difference that exists between postwar and prewar gold standards
and monetary mechanisms in general, it must not be overstressed.1
After all, during most of the period under discussion, a gold standard
was in force in our three countries which in important respects, though
not in all, actually worked much as its predecessor had done and was
managed in accordance with principles and aims that did not differ
toto caelo from prewar practice. This will presently become apparent.
Just now it should be noticed that in this country, though it could no
doubt have done very well without any gold at all, expansion of the
sphere of money and credit was facilitated by the influx of gold (plus
domestic production, which, including that of the Philippines, had been
4.9 million ounces fine in 1915, 2.5 millions in 1923, and then continued
to decline to 2.2 in 1929), in spite of the facts that the monetary gold
stock (gold in circulation plus gold in the treasury and the federal reserve
1 In the first draft of this book it was that difference that was stressed, but the day of
that task has passed. Any truth that fights against traditional habits of mind is in danger
of being overstressed. It is, therefore, not otiose to insist on the following distinctions.
First, if nothing had changed, either in the monetary systems or in the economic and
institutional pattern of the world, we might still have acquired a different attitude toward
the gold standard. Objectively, the gold standard of the twenties might not have been
different, yet thinking might have made it so. This triviality is not superfluous, because
some authors did, in fact, confuse new views with new realities. Second, the identical
gold standard would have worked differently in the postwar environment, e.g., under the
influence of rigidities, barriers, political payments, and so on. This difference is really
the most important one, but whatever consequences it may have had for the "classical"
practice, it was not in itself a reason to abandon "classical" theory. Even of "classical"
practice stronger traces can be observed — no matter whether this was good or bad — if we
cease to libel it by defining it as a set of rigid rules without relation to the general state of the
economic body and if we disregard phraseology (which however was not very novel either) ,
Third, the rules of the game themselves may have been changed. They were changed,
but not anything like as much as is now commonly believed. The great change came in
the thirties.
888 BUSINESS CYCLES
banks minus earmarked gold) only increased to 1927 (annual average
of daily figures for that year, $4.6 billions; 1922, 3.8) and that there
always was "free gold."1 For without that influx — and the free gold —
the Federal Reserve System would have had to steer much closer to the
wind2 and either a problem of protecting the gold value of the dollar would
have presented, or the necessity of some devaluation would have imposed
itself. As it was, the abnormal influx exerted its normal influence, which
was, to be sure, "managed" or regulated but not more so than — in
countries with central banks — it had been in the last prewar* decades.3
An outflow of gold which under more normal conditions might have
persisted, turned into an inflow in the last three quarters of 1920. This
inflow lasted without interruption and in spite of capital export until
December 1924, and then resumed in 1926 and again in August 1928, to
last until October 1929, and to set in again in 1930. Agreement between
this movement in the monetary gold stock and the movement in total
deposits is not much worse than it had been of old. Persistence of "clas-
sical" relations is clearly suggested.
b. The mechanics of what is usually referred to as Federal Reserve
Credit may be put into two graphic nutshells.4 We will first inspect
Chart LVIII.
Reserve bank operations, as recorded by the Combined Balance
Sheet of All Federal Reserve Banks, may be summarized in terms of
1 On this and cognate points see S. E. Harris, Twenty Years of F. R. Policy, in par-
ticular, vol. II, Appendix B, chart on p. 761.
2 Among other things it could not have so readily extended help to foreign governments
and central banks, though most of these transactions, especially the 200 million dollars
credit to the Bank of England granted in the spring of 1925, achieved their ends without
being used. Another form of help may be seen in discount and open-market policies
which tended to stem the inflow of gold. This subject, which we can barely touch, has
been ably discussed by A. Goldstein, International Aspects of Federal Reserve Policy,
Review of Economic Statistics for August 1935.
3 The much-discussed "sterilization" of gold, preventing "gold inflation" during the
twenties, thus reduces to the fact that technically there was always some unutilized margin
of gold. See below. The economists who made so much of that sterilization seem to
have had an inadequate idea of the conditions under which the system worked and which
made that margin necessary. Perhaps they also were victims of a belief to the effect
that prices must rise in a country into which gold is flowing and that only sterilization can
explain their failure to do so.
4 Those two graphs compress, of course, a large amount of fact. They have been
prepared by Dr. Carl E. Thomas, who not only worked up the material but developed the
analysis of this subsection, much of which is taken verbatim from his report. For the
sake of convenience, the description (see also Appendix) extends to 1937. It may seem
hazardous to enter into a subject to which justice cannot be done. But as far as it goes,
our argument is self-contained. For the rest, the reader is referred to the well-known
standard works of Burgess, Reserve Banks and Money Market; Hardy, Credit Policies
of the Federal Reserve System; Harris, Twenty Years of Federal Reserve Policy; and
Riefler, Money Rates and Money Markets.
1919-1929
889
RESERVE ACCOUNT OF MEMBER BANKS CALCULATED
1*2*3*4-5*6-7
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1950 1931 1932 ISB 1934 1935 I9J6
CHART LVIII.— (See Appendix, p. 1076).
890 BUSINESS CYCLES
eight accounts and of certain statistical relations which appear to exist
between them.1 These accounts are:
I. Total Reserves:
Gold, United States Gold Certificates (a small amount), Other Cash
(including silver, a small amount).
Gold dominates this account. In 1919, gold constituted approxi-
mately 97 per cent, and in 1926 approximately 95 per cent of the total.2
II. United States Securities:
This account records changes in the holdings of Government Securi-
ties, i.e., the bulk of Open-market Operations.
III. Bills Bought:
This account records bills acquired by the reserve banks. Bankers*
acceptances are included under this title. During the period under
review, bills were often purchased or sold to offset seasonal variations
in the demand for currency and other central-market disturbances of a
routine nature, such as government fiscal operations. To this extent
these transactions are properly looked upon as open-market operations.
But they really constitute a case intermediate between II and VI, because
the initiative often came from the member banks.
IV. Float.3
V. Notes in Circulation:
Federal reserve notes and Federal reserve bank notes issued. This
includes Notes of Other Reserve Banks, the amount of which is, however,
1 Account totals are taken as of the "end of the month" in all cases. When figures
are actually reported for the last day of the month, these have been used. When reports
are available as of the end of each week, the report of the last week in the month has been
used. Although "average daily" figures are available for all the major accounts, these
data could not be used, since they are not in sufficient detail to enable checks upon content.
2 Prior to January 1984, gold was valued, in the Combined Balance Sheet, at $20.67 per
fine ounce. After that date it was valued at $35 per ounce. The gold account and the gold
itself was then, at the old value of $20.67, transferred to the Treasury. In its place
appeared a new account: "Due from the U. S. Treasury." Changes in this account, after
January 1934, are recorded at the rate of $35 per ounce.
8 Before January 1926, this account is represented by the difference between Uncol-
lected Items and Deferred Availability. Its movements were insignificant after 1921.
After December 1925, equal and offsetting amounts of "Uncollected Items" and "Deferred
Availability" appear upon the Reserve Bank Statement. The previous difference is
recorded in two separate accounts, namely, Float and Federal Reserve Notes of Other
Reserve Banks. It is clear that previously notes of other reserve banks were treated as
an uncollected item. Since they form a part of Notes in Circulation, it is necessary to
add this item to the Float as published, to obtain a figure comparable to the previously
recorded difference between Uncollected Items and Deferred Availability.
1919-1929 891
recorded for a period sufficiently long to warrant the assumption that it
is not significant, so that the item may be roughly equaled to notes in
outside circulation.
VI. Bills Discounted:
Member banks' rediscounts or indebtedness.
VII. Government Deposits:
This account was negligible until recently.1
VIII. Member Bank Reserve Account:
This account, often considered by practical bankers as the member
bank clearing account with the reserve banks, records the indebtedness
of the reserve banks to member banks.
It will be seen that the algebraic sum of accounts I- VII produces a
time countour which is practically identical with that of VIII, the member
bank reserve account. The difference in level between the two represents
the net sum of All Other Accounts appearing on the balance sheet of the
reserve banks.2 This sum exhibits practically no current fluctuations.
Its amount was in the neighborhood of 300 millions.
Now we subtract Notes in Circulation from Total Reserves plus
United States Securities plus Bills Bought plus Float, and find an almost
perfect inverse and linear relation between the result and Bills Dis-
counted. See Chart LIX.
In a statistical — and a short-time — sense it may be said that members'
indebtedness is largely "explained" by the net of those Five Accounts.
And so is, hence, the reserve account, which is a function of this net and
its relation to rediscounts or indebtedness.3 It follows, in confirmation
of what has been read earlier in this chapter, that members in the aggre-
gate did not habitually borrow in order to expand their operations.
They borrowed primarily in order to avoid having to contract them,
i.e., in order to replenish their reserve accounts when the net of the Five
1 In the Bulletin of the Reserve Board, published in January 1936, it is definitely stated
that the recent increase in Government Deposits was the result of the purchase of Govern-
ment Bonds by Member Banks for which they surrendered Reserve Balances. This
method of government finance involves a departure in technique from previous procedure.
2 All Other Accounts consist of: Non-reserve Cash, Other Securities, Foreign Loans
on Gold, Due from Foreign Banks, Bank Premises, All Other Resources, 5 per cent Redemp-
tion Fund, Foreign Bank Deposits, Other Deposits, Capital Paid in, Surplus, All Other
Liabilities, Reserve for Taxes.
3 A qualification is necessary with respect to government deposits. When they increase
at the expense of members' reserves which, as pointed out above, has been sometimes the
case in the last years, such increases would have to be added to the latter, in order to get
the right contour. For the period under survey, however, this has no importance.
892
BUSINESS CYCLES
A -TOTAL RESERVES
fl»A PLUS US. SECURITfES
C-B PLUS BILLS BOUGHT
0-C PLUS FLOAT
E= D MINUS NOTES IN CIRCULATION
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936
CHART LIX. — (See Appendix, p. 1077.)
1919-1929 893
Accounts decreased. And when that net increased, that is to say, when
they experienced an access of funds, they did not primarily expand
operations. They first of all reduced their debt. In some cases com-
pensation was exact, or almost so. From November 1924 to March of
1925, for instance, the federal reserve banks sold 260 millions' worth of
governments and bought 75 millions* worth of acceptances. The differ-
ence was practically made up by members' borrowing, which amounted
to roughly 175 millions. In other cases the compensation was not nearly
exact. We are not holding that expansion of loans and deposits was at
no time accompanied by an uncompensated surplus of borrowing, still
less that increase in our net was wholly compensated by repayment.
The rediscounts, nevertheless, stand out as the variable primarily
affected by the net and very little affected by anything else — a fact
which is very important for any theory of the American central market.
Two other relations have been exhibited in the chart which illustrate
what fundamentally is the same point, and characterize the same mecha-
nism. First, the close short-run covariation, but slightly disturbed in
1929, of open-market rates, represented by the commercial paper rate,
and rediscounts. Nothing can be more indicative of member banks'
reluctance to being in debt than this fact, which is due to the promptness
noticed before with which they contracted open-market commitments
in that case. Or, to put it the other way round, the inverse relation
between open-market rates and the net of the Five Accounts indicates
practically complete short-run dependence of the former on the ebb and
flow of "funds": as a matter of surface mechanism and in the short run,
variations in these "funds," consisting in member banks' cash plus
reserve balances, were the dominant factor in the variations of their
debts to reserve banks, and variations in these debts the dominant factor
in the variations of interest rates which, and practically nothing else,
were, hence, amenable to control by the regulation of that ebb and flow.
Now, putting aside internal drains and remigrations of cash, which, as
we have seen, were of but secondary importance during the period under
survey (at least after 1923), we may say that the variations of members'
"funds" were a function of central creation and of gold movements.
Central creation proceeded from the initiative of members (rediscounts;
hereafter called Responsive Central Creation) or from the initiative of
federal reserve banks (open-market operations in governments and to
some extent also in acceptances, though these constitute an intermediate
case — see above — hereafter called Autonomous Central Creation), only
the latter being now included in that ebb and flow.
Finally, gold movements, whatever their cause, acted in the same way
as autonomous central creation and must hence be, in this connection,
algebraically added to it. But if we take them separately and if we
894 BUSINESS CYCLES
choose intervals of time short enough, we find that they are positively
related to open-market rates with a decreasing lag.1 On the chart,
short-run gold movements have been represented by month-to-month
changes in total reserves (which were mainly gold — see above), and
open-market rates, by the yield on government certificates (but any
open-market rate would do). This completes our picture by suggesting
that total reserves, one of the chief determinants of short-run variations
in open-market rates, are in turn themselves influenced by the latter,
and thus supplying a link in a general relation of interdependence.
Very familiar traits (see Chap. XIII) of prewar patterns thus emerge
in the postwar picture. Whatever theories we may make our own, facts
behaved and mechanisms worked, during the period under survey, in a
manner that was anything but revolutionary.
c. It remains to sum up the teaching of our charts from the standpoint
of the Federal Reserve System and to appraise the effects of its policy
on the course of cyclical phases. To formulate this policy in terms of
intentions would be an impossible task; for, though this method is in
any case hazardous, it becomes impracticable in the case of so acephalous
a body as the one before us. A central organ of banking would, under
the conditions and in the mentality of this country, always have an uphill
fight against the public, the business community, speculators, senatorial
farmers' friends, member banks, and hence find it very hard to acquire
enough authority to pursue any policy consistently. But what or who was
the central organ ? There were, first, the reserve banks themselves, with
which from the outset a considerable amount of initiative and autonomy
had been lodged by the Federal Reserve Act, but which soon acquired
a sort of customary right to initiate or " propose " the general policies of
the system. Second, the Federal Reserve Bank of New York must be
listed as a special "reserve authority," because its prominent position,
its European connections, and its particular interests gave it a power
and a slant quite out of line with those of the others, and because it
evidently aspired to, and to some extent conquered, under the leadership
of a strong man, the role of "the" central bank. Intimately connected
with the New York bank was, third, the committee on open-market
operations, which eventually developed into the most active element of
the organism. Fourth, there was the Treasury, which had by no means
forgotten its ancient role in the money market. And, fifth, there was
the Board, which, sometimes at war with one or more of those other
organs of general banking policy, at first tried to assert such powers as
had been vested in it but soon drifted into the position of a coordinating
1 The lag seems to have stabilized itself at about 4 months. After 1930 the relation
ceases to hold. But many factors in that disturbed time — •" capital flight" from Europe,
for instance — readily supply explanation.
1919-1929 895
agency. This role, however, it filled with astonishing success, in spite
of the fact (or because of it?) that there are unmistakable symptoms
pointing to a serious division of opinions within it, which often paralyzed
decision. The writer understands that election to a seat on the board
was a coveted honor. But the seats must have been extremely uncom-
fortable ones, even if economists of all nations had not completed vexa-
tion by discovering that the Federal Reserve System was an entirely
new departure, harboring unheard-of possibilities, sure to put an end to
the recurrence of depressions, etc., and by thus raising, even among other-
wise sane and intelligent people, hopes that were as unwarranted as their
disappointment was certain to be attributed, in self-defense of the
enthusiasts of scientific and extrascientific description, to the incompe-
tence of the board.1
But fortunately we need not trouble about intentions and phraseo-
logies. The actual behavior which the logic of the situation soon shaped
into a definite pattern — 1922—1923 roughly dates the decisive steps in
its evolution — suffices for our purpose. Before that date, a tendency
to help in the process of normalization after the war, as evidenced, for
example, by the circular which reminded banks that war credits would
not indefinitely be prolonged, is the only thing to note. Criticism of this
attitude and the discovery that those outlets for the reserve "funds,"
which the Federal Reserve Act had contemplated, would not develop —
i.e., the business man's typical mortification at the absence of adequate
earnings — perhaps provided the first impulse to embark upon buying
governments at the beginning of 1922.
Turning again to Chart LIX, we are immediately struck by the fact
that no energetic use was ever made of the discount rate. Not only was
it always kept below even the commercial paper rate — thus never occupy-
ing the position or filling the function of German or English bank rates —
but it also followed the market in every single instance except — and
1 That body was so unfortunate as to draw fire even from the other party to the
Battle of "Theories." Economists who did not share those hopes and quite correctly
appraised the limitations of the doctrines on which they were based, are or were in the
habit of accusing the board of misdirected measures responsible for what most of those
economists would call inflation and, partly or wholly, for the subsequent crash. There
are two mistakes in this, apart from the incorrect allocation of responsibility to the Board.
First, such criticism seems to imply that these economists yield to their opponents to the
extent of sharing part of the latter's exaggerated ideas of the importance of what central
bank policy can make or mar. Second, they do not seem to take adequate account of the
structure of the American financial engine and of the data confronting the board from
the outset. The present writer is not offering laurels to the board. If he were hunting for the
most appropriate sculptural ornament for the board's building, a statue of Hercules would
certainly not be among the first to occur to him. But critics do make their task a little
easy. It should be added that our list of "central organs" is by no means complete; but it
serves to convey the essential point.
BUSINESS CYCLES
even this exception is more apparent than real — at the very beginning
of 1928, when an effort was made "to gain control." But at the same
time it follows from our analysis that there was another reason for this
besides the "political impossibility" of handling that weapon boldly: to
some extent the Federal Reserve System in fact controlled the situation
of which market rates were the outcome and symptom, so that in fol-
lowing or ratifying market by bank rates it really followed and ratified
the result of its own action — bank rate as an independent weapon had,
in fact, practically ceased to exist. Such leadership or management as
there was, was exerted through the Five Accounts, mainly by open-
market operations, supplemented by suasion. As we have seen, this
method requires in order to be effective a certain behavior on the part
of members, in particular a certain attitude toward being indebted
to reserve banks, for if they had all stormed for more credit and for
permission to use it permanently and as a matter of course, they would
have had the public behind them almost to a man, and any resistance
by the board or the reserve banks would have been swept away. But
this kind of discipline the better class member banks displayed through-
out, and the greatest achievement of the board was its success in training
them up to it and in establishing a professional tradition which made it
derogatory to a bank's standing to be in the red at its reserve correspond-
ent, except in order to tide over temporary pressure, or, for a first-class
bank, to be in the red at all, even when the reserve system seemed to
invite it by low rates. A qualification has to be added, however. We
have said that discipline was kept unbroken until the crisis. This is
true, for the whole period, only in a formal sense. In the spirit it holds
only to the beginning of 1928. Then many important banks, especially
New York banks, kicked over the traces by creating acceptances and
selling them to reserve banks, not so much in order to evade the higher
discount rate — this was a by-product — as in order to evade the necessity
of running deeper into debt. The reserve banks bought, and were by
the board allowed to buy, all that paper without demur. This is the
most important instance of a practice which induced us above to qualify
our inclusion of purchases of acceptances in open-market operations.
Barring this, the mechanism functioned. Although, as has been
pointed out, open-market operations did not directly increase and decrease
the borrowing facilities available to nonbank firms because the member
banks (largely) compensated them — if, indeed, they were not from the
outset intended to compensate, hence to be compensated by, gold move-
ments— the policy was not thereby defeated since, whenever open-market
operations decreased the net of the Five Accounts, banks were forced
to borrow or — which served still better — to reduce their open-market
commitments. But the corollary to this is that open-market sales never
1919-1929 897
went further than this, t.e., beyond the compensating powers of banks.
Perhaps they were not intended to. The committee on open-market
operations had every reason to fear the reaction if its measures had
touched "legitimate business." In any case, the effects which can
be produced by this weapon, unless reinforced by simultaneous refusal
to buy acceptances and by rationing credit to member banks, are limited
to the available ammunition, which it may be impossible to collect.
The reserve banks' annual average of United States securities holdings
in 1919 Was 261 millions — no great impression could have been made
by any sale from this if there had been any wish to influence in this way
the situation then prevailing. Luckily, the first open-market operation
to be undertaken with a therapeutic intention was a purchase and not a
sale: from January to May 1922, about 400 millions worth of governments
was bought on the top of a steady influx of gold. In a sense this was the
decisive step that raised the Five Accounts to the level under which
they were, except in 1923, never again allowed to fall, and thus per-
petuated the plethora of money. The alternative would have been
to enforce the liquidation of member bank indebtedness without pro-
viding the funds for it. This would have spelled pressure, slower
recovery, sobered advance, much less speculation, and, after 1929, a
milder depression.
No real pressure was exerted by the next open-market operation,
which was the first to offset a gold influx and consisted in the sale of
about 525 millions United States securities between June 1922 and July
1923. Selling actually was, while it lasted, at a greater rate than gold
influx, so that member banks were for the time being forced to borrow,
but the continuing stream of gold alone soon made up for this and
even during that time effects were amply compensated by gold plus bills
bought plus borrowing, so that they did not extend beyond the central
market. It should also be observed, in order to get at a sound appraisal
of the achievements of this policy, that, as far as the preceding open-
market purchases and the gold influx, as well as the incident reduction
of member banks' indebtedness by over 600 millions, did speed up
recovery — and while we deny that it mode recovery and hold that it was
all but futile as a remedy for depression, we do admit that the temperature
of positive phases can be raised by central bank policies — they also helped
to bring about the situation which the board or the reserve banks then
felt to stand in need of restrictive correction. Thus a current view on
the success of this policy requires to be corrected on two distinct heads :
first, success in influencing cyclical situations was much smaller than it
looked to post hoc ergo propter hoc analysis; second, such success as there
was, in part consisted in correcting the effects of the reserve system's own
policy,
898 BUSINESS CYCLES
What we might call the first conjuring trick followed. In the Kitchin
recession and depression of 1928—1924, when finance and business felt
a vague malaise after their doings in 1922-1923, the reserve system stepped
forward to chase away darkness, to insure stability of prices and to
guarantee new-era prosperity by buying about 510 millions' worth of
governments between December 1923 and September 1924, which, in the
minds of some economists, was simply equivalent to increasing active1
deposits by 5,100 millions. In reality, it did not mean this but deposits
did react, as well as member banks' debts, and the incident fall in open-
market and reserve bank rates turned the tide of gold. The latter
effect was really the most important part of the success, inasmuch as,
partly because of European complaints and partly because of a percep-
tion of the difficulties which continued influx of gold might eventually
produce, the turning of the tide was one of the objects of the measure.2
As soon as November 1924, selling was again resorted to and continued
to March 1925, when the total amount reached was 260 millions. This
operation has already been mentioned in order to illustrate the modus
operandi. Effects were as before, but the operation deserves to be
emphasized as a token and measure of the success with which, braving
many difficulties, the reserve system kept its hand on the steering wheel
and actually followed a definite course. And the same may be averred
about the two smaller operations, purchases in the amount of 65 millions
in April 1926, and sales in the amount of 80 millions in August and
September of the same year.
A situation closely similar to that of 1923-1924 — the similarity
extends to the presence of European influence, this time accentuated by
a pilgrimage to the miraculous shrine — induced another purchasing
campaign in May 1927 which lasted to December and raised the United
States securities holdings of the reserve system by about 300 millions or
to about double their previous figure. There was by then nothing unusual
either in the occasion or the effects or the amount of the operations. If
we did not know the contrary, we should not from the facts of the case
infer that any controversy would have arisen about this particular opera-
tion or that the latter would have been considered as particularly admir-
able by some and as particularly mistaken by others, or finally that
1 Time deposits in reporting member banks, however, increased during 1924 by nearly
% billion.
2 As far as it was, the purchasing operations may have commended themselves to
members of the board who neither believed in their efficacy as a remedy for the relapse
in business nor would have approved of using the tool at that time, even if they had believed
in its efficacy. Thus, as happens so frequently in all spheres of life, men were able, from
entirely different premisses and with different and conflicting objects in mind, to agree on
the particular measure.
1919-1929 899
praise or blame for it should have been fastened upon a particular man.1
The reason for all this is, of course, in the fact that that was the last
buying operation before the stock exchange crash of 1929. We fully
share this motive, since we have had to stress both in our historical nar-
rative and in our time-series analysis many an abnormality2 in the eco-
nomic processes of 1928 and 1929. But no amount of careful searching
establishes the connection one has a perfect right to suspect. Such a
search reveals indeed many individual facts which point to the presence
of "artificial stimulation." For instance, gold exports were more than
offset, and members' investments, total deposits, and — what is particu-
larly significant — brokers' loans for own account of member banks, not
only on account of others, were increased. But the effects of all this
may, as we know, easily be exaggerated, and nothing really adequate for
explaining either the 'boom of 1928-1929 or the subsequent^ breakdown
can be linked to that buying operation per se as the sole or as a major
cause.
In fact, there is good reason to doubt whether it played any role in
the causal pattern, for the reserve system acted with more than usual
vigor to stamp out any sparks that its action may have set flying about.
First, mere cessation of buying had, because of the outflow of gold, some
restrictive effect as early as December 1927.3 And in January 1928 the
reserve system embarked upon the biggest selling operation but one
of its history, in which it persisted fully as long as was compatible with
safety, i.e., to April 1929, when sales summed up to 405 millions and
total open-market operations since January 1922, as already stated, to
minus 65 millions. Simultaneously bank rates were repeatedly raised,
eventually (July 1928) to 5 per cent,4 member bank's indebtedness (which
however, as we have seen above, did not in this case quite fulfill its
function) passed the billion line, below which it had moved ever since
1922,5 and brokers' loans for own account dropped below 1 billion, while
gold exports continued through July. This seems to have been drastic
action if ever central bank action was. Nor was it futile, as some, or
1 Mr. B. Strong's influence may have been much in evidence on this occasion. But
the board certainly did not offer significant resistance, for thtey obligingly brought the
reserve bank of Chicago into line when it displayed reluctance to reduce its rate along with
the others. Nothing would have been easier than to checkmate Mr. Strong by letting
Chicago have its way, which would have tightened the New York market.
2 They cannot be listed again, but if the reader wishes to get the argument in full he
should now go back and list them for himself.
3 It is, however, true that the Treasury's needs for funds arising out of the conversion
of the second Liberty Loan, which induced it to take and partly use an overdraft at the
Federal reserve banks, may have counterbalanced that effect temporarily. But the net
efflux of gold was 67.42 millions in December alone.
4 The last step, from 4.5 to 5, was, however, taken by only eight reserve banks.
6 The monthly average for 1920 had been nearly 2.6 billions and that for 1921 over 1.7.
900 BUSINESS CYCLES
oppressive, as other people thought. It tightened the central market.
It did not exert pressure on industrial and commercial business, as is
clear from the simultaneous increase in customers' loans which has been
noticed before. And it was quite rationally discontinued in the second
half of 1928, when the gold influx, which resumed in August, was allowed
to take effect.
If the board needed any lesson, it was supplied by this experience,
which conclusively showed that (relatively) high rates will attract gold,
that for this and other reasons no impression could, by ordinary methods
applied within ordinary limits, be made — except very temporarily (middle
of 1928) — on the stock exchange, and that the most immediately danger-
ous breach in the wall was the loans on account of others. It is sub-
mitted that the inferences which might have been drawn from this lesson
are, first, that bank rates should be lowered1 and, second, that loans
to the stock exchange should be attacked directly, though the lowering
of bank rates through turning, stopping, or reducing the flow of gold
would already have done part of the job. The first was not done, per-
haps because it seemed too unorthodox a thing to do. But the second
was exactly what, obviously taking its courage in both hands and drop-
ping its attitude of dignified reserve, the board actually did2 — or tried
to do. It announced its intention, in the famous circular that was
published on Feb. 7, 1929, "to restrain the use, either directly or indi-
rectly, of federal reserve credit facilities in aid of the growth of specula-
tive credit/' The announcement was embarrassed, overcautious, and
also not quite sound in explicit and implied argument, but in essence it
proclaimed Mr. Schacht's policy of 1927, i.e., the policy of forcing banks
to withdraw their loans to brokers, whereby other lenders would be
induced to retreat. No doubt is possible that this measure would
have been quite successful if resorted to earlier, say, by one or, still better,
two years. Then it would have prevented, whereas at the beginning of
1929 it could only have precipitated, the crash. The perception of this
accounts both for the determined opposition — led by the reserve bank of
New York3 — and for the readiness of the board to yield, in June, after its
1 We thus find ourselves pro tanto in agreement with many fellow economists, among
them Professor Irving Fisher and Dr. Currie, with whose arguments and results we do not
otherwise agree. It will, however, be realized that we arrive at the same result from very
different premisses and expecting quite different consequences.
2 There are many methods for achieving the same result, taxation amounting to con-
fiscation of gains from stock speculation and others, but the one chosen was presumably
the only one open to the board.
3 That bank followed up its previous performance in the field of central bank policy
by the proposal, made a few days after the board's announcement of direct action, to
raise the rediscount rate instead. If the above analysis is correct, this, through attracting
additional gold, would have given another impulse to speculation The reader will recall
a similar observation previously made in our discussion of a similar situation in Germany.
1919-1929 901
policy had fully demonstrated its effectiveness by putting a stop to the
rise in stock prices and reducing brokers' loans by between 600 and
700 millions. By then, however, it mattered little whether it yielded
or not.
The reader will please formulate for himself whatever may seem to
him to follow from this analysis as to the merits of the actual practice
described or of the various alternatives advocated at the time and later.
For the purposes of this book it is sufficient to state or restate three
conclusions to which we have been working up all along.
First, whatever conflict of "theories" and intentions there may have
been, we have seen that the actual behavior of the organs of central
banking in this country reveals a very definite pattern. This pattern
greatly differs, of course, from that which we have tried to piece together
for prewar times from the behavior of such central organs as the
American banking system then possessed, since these had little facility
for concerted action. But it differs much less, in essentials as dis-
tinguished from forms and methods, from the prewar patterns of central
banking in countries which then had individual central banks. Gold
movements, in particular, and the state of gold reserves indeed did not
play the role which it was usual to assign to them in prewar theories of
central banking. But this was partly due to exceptionally favorable
circumstances which would also have pushed reserve considerations into
the background for any prewar central bank; and though the effects of
gold movements were no doubt regulated, as they had always been, they
were not therefore obliterated. Open-market operations were no
novelty, nor did they serve in fundamentally new ways or for fundamen-
tally new purposes. American conditions put them into the limelight
and enforced their systematic use, but that was all.
Second, it follows from this, together with our general view on the
role and possibilities of central banking, that the cyclical processes of the
period were not substantially affected by the policy of the reserve system.
Its attitude to the ups and downs of business was much the same as the
attitudes of central banks had been for decades before the war. It
"regulated" the central market and, doing this ably and conscientiously,
achieved what can, and failed to achieve what cannot, be achieved in this
way and what equally ably and conscientiously managed central banks
achieved and failed to achieve before. The successes of 1924 and 1927
and the "failure to prevent the great depression" are largely figments of
the mistaken theories of the day. With respect to that failure, however,
a secondary though still important qualification is necessary.
Third, this qualification bears upon the well-known controversy
whether or not there was "inflation" or "deflation" and whether or not
the reserve system and its policies were responsible for it. Couched in
902 BUSINESS CYCLES
less ambiguous terms, the question reads whether there are in the behavior
of time series any deviations from expectation which are traceable to acts
or attitudes of the reserve system or other autonomous monetary factors.
All the evidence presented points to the conclusion that, so far as fluctua-
tions and "trends" were concerned, this was not the case until the spring
of 1928. Quantities, prices, values, incomes, and so on behaved sub-
stantially as we should have expected them to behave in the absence of
disturbance from the monetary sphere, although monetary magnitudes
and expressions moved on a level explainable only by the uifcorrected
monetary disturbance caused by the war and by war finance. The
attempts that have been made to allocate responsibility for the occur-
rence of the great depression to monetary and banking factors, or to the
behavior of economic elements primarily shaped by them, sound as
unconvincing if they come from authors with "inflationist," as they do
if they come from authors with "deflationist" sympathies. In particu-
lar, there is little more than assertion in the theories which hold that
interest rates or prices were either too high or too low or declining too
slowly or too rapidly. Our methods are rough. We must, hence, admit
the possibility and even likelihood that there may be elements of truth
in all those and similar assertions, even in mutually contradictory ones.
But they are obviously inadequate to the phenomenon to be explained:
common sense and common experience should be sufficient to convince us
that if explained by maladjustments such as these, the great depression
remains unexplained. Reserve policy in particular seems rather to have
warded off than to have caused disturbance, rather to have eased than
to have impeded adaptation, rather to have brought out than to have
destroyed or controlled the essential (expected) contours of the cyclical
process — the last point being illustrated, for instance, by the fact that
rediscounts moved so well in the cyclical phases. The paradox that has
been so keenly felt by economists and noneconomists, viz., how such a
thing as the great depression could have occurred if, though not every-
thing, yet so much in the economic organism was "quite all right,"
obviously calls for, and is resolved by, the more fundamental explanation
which the logic of the capitalist process supplies.
But if that is so with respect to the occurrence of the depression, it is
not necessarily so with respect to its intensity. And the latter links
up with the breakdown of stock markets and this again with the abnor-
malities of the last 2 years of the period. Then it was that things got
out of hand, in spite of the fact that the reserve system used its regulating
powers fully as much as the nature of these powers seemed to permit.
But what precisely were the things that got out of hand ? The immediate
trouble was with the stock market and with loans on account of others.
We need only follow up this clue in order to get an answer which will
1919-1929 903
supply also the qualifications to be added to the above statements about
reserve policy and its effects. The loans on account of others were but
a symptom of the abundance of money that prevailed throughout and of
the powers of the deposit manufacturing engine. The increase in loans
and investment of All Banks which occurred from 1922 to 1929 amounted
to about (difference of annual figures) 18.5 billions and tells but part of
the tale. The other part is in the fact that at any given moment
funds in existence were much above business requirements, the excess
being uncontrolled and uncontrollable. So long as, under the influence
of the lesson of 1921, their owners and the banks kept discipline it was
possible for the reserve system to maneuver successfully. As soon as
they ceased to follow the lead, as they were bound to do sooner or later,
the consequences followed automatically. We are not going to discuss
responsibilities or the question whether the diiignosis should be expressed
by saying that there* was potential "inflation" all along, which turned into
actual "inflation" in 1928. It is enough that this course of events,
whatever its appropriate name, could, though not by routine methods,
have been prevented in 1922 without anyone except speculators being
the worse for it — even at the time — and that this necessarily would have
made a great difference in the intensity of the depression.
d. The policy of the Bank of England and the conditions in the
English open market may be described as they have been above (Sec. C)
in terms of the effort made, first with increasing, then with decreasing
success, to maneuver back to the gold standard at prewar parity and to
keep on it. As has also been observed already, this did mean pressure
on the economic process. And gold movements and states of the gold
reserve, which was no longer protected by an unchallenged world-banker
position as it had been in prewar days — no longer, especially, by a thick
wall of well-disciplined and quick short-term assets in foreign countries —
hence also money rates, were more important and more delicate to handle
than ever. The central bank aspects of this are obvious from Chart LX.
That situation, complicated as it was by the practical necessity of
resuming long-term foreign lending, especially within the empire,1 was
as difficult in practice as it was unproblematical in theory. If the policy
of the Bank of England had not been given such undue prominence,
both in the analysis of the shorter fluctuations of English business and
employment and in the causation of the great depression,2 no further
comment would be necessary. As it is, we must stay for a moment.
The Bank of England encountered, within its general policy, which
was dominated by the exigencies of that maneuvering, current business
situations in exactly the same spirit and in much the same way as it
had before the war. In 1920 it acted with less promptness and went
less far than on some previous occasions, but still raised its rate to 7 per
904
BUSINESS CYCLES
cent on Apr. 15 and then took 22 months to return (Feb. 16, 1922) to 4.5.
However, the gloom of English business prospects, which could not have
been relieved by money at 2 per cent, and the brightness of American
PRO PORT ON
DEPOSITS AMD BA^K PbST ftlLLJ
\
tOTES
LONCON HANK
PHER'
f ROPO TTIOh
JALA slCES
B^LANCESTC RESfRVE
\
LOhDON
\
BANKERS'
NOTE
HER"
V
r~\
SECU ^ITIE
OF P JBLIC
\
1920 1925 1930
CHART LX. — Bank of England series (see Appendix, p. 1077).
prospects, which could not have been darkened by money at 7 per cent,
obviously offer a much more plausible explanation of the difference in the
1 The relative importance of loans to the Dominions or to borrowers in the Dominions
greatly increased in the postwar decade, though total long-term foreign lending never
came near the 1913 figure (197.6 million pounds). Loans to borrowers in the British
possessions amounted (in million pounds) to 69.2 in 1925, 57.5 in 1926, 98.2 in 1927, 81.5 in
1928. The balance of payments for the same years was: +54, +9, -j-114, +149.
2 The chief authority to sponsor what seems to the writer a seriously wrong view is
Mr. Hawtrey. C/., for example, Trade Depression and the Way Out, new ed., 1938, pp.
20 et «<?<7., where he presents a sketch of English postwar business fluctuations viewed
almost entirely as functions of bank rate, which in turn is made a function of the absorption
of gold by France and the United States. Since the bank, considering the precarious
position of the pound, had to be very cautious in following its traditional policy of lowering
the rate in negative phases and to keep it low in recovery, and very prompt in raising it in
prosperity, it is not difficult to establish correlation or lagged correlation. But we know
that this proves absolutely nothing for the presence of a cause-effect relation.
1919-1929 905
speed of recovery in the two countries. After that, cheap money policy
prevailed for a time, as it had always done after the passing of a "deep"
depression, though, under the circumstances, it was not possible to carry
it quite so far as it had been carried in the comparable segment of the
preceding Kondratieff.1 For the next 3 years, i.e., before the pound
reached parity, the rate was on the average actually lower than after,
which is very natural because maneuvering up to parity was a much
easier task than maneuvering to keep on it. The latter had to be done,
not only without the help of international speculation, but in the face
of constant danger of attacks from it — for selling short was thenceforth
(beyond costs) riskless. In the domestic situation there was, in fact,
nothing to prevent the Bank from going on with the rate of 3 per
cent, which had been reached on July 13, 1922, and retained to July 5, 1923,
although the Federal Reserve System sold United States securities2
during that year. But eventually the strain proved too much and on
July 5, 1923, the Bank's rate went to 4 per cent, although the Federal
Reserve System then discontinued sales. It remained there until Mar. 5,
1925, in spite of the Federal Reserve System's buying operations in
1924 — presumably, in order to enhance their effect on the Bank's posi-
tion— and also in spite of the sales that followed them. But then came,
with only a 2 months' interruption, over 2 years at 5 per cent. This rate
was no doubt higher than would have prevailed without the abnormalities
in the situation. But recovery turned into prosperity all right, and if this
term be objected to, the behavior of the index of production is, in any
case, conclusive evidence of a considerable improvement, which, it is
true, was seriously impaired by the great social struggle of 1926. In
1927 the easy money policy in New York was cautiously responded to
by a reduction of bank rate to 4.5 per cent, which, moreover, was not
always made effective. The year 1929, of course, revealed the intrinsic
weakness of the situation. The increase of the bank rate to 6.5 per cent,
on the surface not more serious than its apparently similarly conditioned
peak in 1899, and the decline of the Bank's gold stock below the Cunliffe
level ushered in the death struggle of the pound. But it no more caused
the depression than a similar rate had caused a depression in 1899.
1 If 1865 to 1874 be accepted as a roughly comparable segment, we may note that the
monthly fluctuations of the average for those 10 years were between a maximum of 4 pounds,
8 shillings, 8 pence (November) and a minimum of 3 pounds, 2 shillings, 2 pence (Septem-
ber). For 1875 to 1884 the monthly maximum of the 10-year average was between 3
pounds, 18 shillings, 2 pence (November) and 2 pounds, 15 shillings, 8 pence (April). See
Palgrave, Bank Rate and Money Market, p. 97.
2 For our theory is : when the Federal Reserve System sold, this stiffened, other things
being equal, the New York open market and attracted gold to America, which would also
tend to stiffen the London market and, moreover, weaken the pound and vice versa in
the case of buying. In fact, European influence was, obviously on the same theory,
exerted in favor of buying operations in 1924 and 1927. To some students this will seem
too evident to be worth stating, to others it will seem wrong.
CHAPTER XV
The World Crisis and After1
A. The World Crisis and the Cyclical Schema. — We should now be
able to answer the question how far the course of events from the fall of
1929 to the summer of 1938 can be described in terms of the analytic
model presented in this book and how far other factors, external or inter-
nal, new or old, must be relied on for explanation. From the standpoint
of this question alone and merely as a last exercise in application of that
model we will review the sequence of situations and some of the policies
of that period. This object is not so modest as it might seem. The
answer will, directly and by implication, economically and sociologically,
cover much more ground than the question at first sight suggests. But
precisely because it will, the reader should be reminded once more that
no conclusion we may arrive at can do more for him than to help him to
see things more clearly. From that to practical evaluations or recom-
mendations the way is long indeed. And everyone of us must, carrying
his own individual load of desires, prejudices, and visions, travel it alone.
Whenever any set of propositions and observations leads us to expect
a certain event, the actual occurrence of that event will always strengthen
our confidence in those propositions. In such cases we are in the habit of
saying — at the risk, as we have had ample opportunity of observing, of
some violence to logic — that they are "verified" by that event or that
they "explain" it. Now in this sense the occurrence at that time of a
severe and prolonged depression in itself verifies or ratifies the application
of our model, i.e., warrants explanation in terms of our process. For we
need only survey, in the light of our interpretation, the developments
since 1898 in order to understand why such a depression should have
occurred as part and parcel of that process. No claims, be it repeated,
1 Although an endeavor will be made at the expense of some repetition to protect the
following exposition from certain misunderstandings to which some statements to be made
are liable if taken by themselves, this chapter should not be read before the analytical
apparatus presented in this book has been fully mastered: the theory of the "world crisis"
begins on page one. No new charts will be presented, but those that have been used in
the preceding chapter cover and illustrate the main points of the argument of this, and
should be referred to again.
906
THE WORLD CRISIS AND AFTER 907
are made for our three-cycle schema except that it is a useful descriptive
or illustrative device. Using it, however, in that capacity, we in fact got
(in Chap. XIV, Sec. E), ex visu of 1929, a "forecast" of a serious depres-
sion embodied in the formula: coincidence of depression phases of all three
cycles. It will be well to recall once more what this formula precisely
means. For reasons we know, capitalist evolution spells disturbance.
We also know that it spells simultaneous disturbances of different order of
importance and different range in time. Junctures therefore occur in
which the symptoms incident to scrapping and rearranging dominate the
scene. Among these junctures there are some in which adjustments to
long-range and more fundamental, and adjustments to short-range
and less fundamental industrial changes do not occur at the same time,
and there are others in which they do. In the first case, symptoms
will be mitigated; in the second, intensified — or, to return to the schema,
in the first case the depressive phases of one or two cycles will hit a
"floor" provided by the nondepressive phases of the other or the two
others; in the second case there is no such floor and hence a more serious
and especially broader trough1 — all of which could easily be translated
into terms of hard business fact.
In XIV, Ewe went a step further by making the — absurd — assumption
of strict periodicity of all the cycles and equal duration of all their phases.
Counting on in the same manner, we should get a Juglar depression — on
the Kondratieff depression which dates from the fall of 1925 — from July
1930 to the middle of November 1932, which should be severe, to be
followed by a Juglar recovery to March 1935, which we should expect to
be, owing to its position within the Kondratieff, slow and weak. That
depression would contain three Kitchin phases, a depression to the middle
of April 1931, a revival to January 1932, and a prosperity to the middle of
November 1932; and the latter two should assert themselves mainly
(not wholly) by decrease in rates of decrease. This schema has none
except illustrative significance and no value attaches to the dates.2
1 That the situation of 1980 answers well to a description in terms of the three-cycle
schema has been noticed by Professor Alvin Hansen, Economic Stabilization in an Unbal-
anced World, 1932, p. 95.
2 But we may still compare at least the general import of our schema with the predic-
tions made by forecasters in 1930. Professor Warren Persons, for instance, (Forecasting
Business Cycles, p. 44) predicted the upturn for February to April 1931. This was by
no means so compromising a failure as people — and Professor Persons himself — thought.
It was arrived at in a perfectly workmanlike way by methods which it was quite possible
to defend. It is a not uninteresting question to ask what the failure was due to. We
observe that Professor Persons' dating corresponds fairly well with ours as to the Kitchin
revival. This is more than a coincidence. He was reasoning on the course of short
cycles — though not exactly of our Kitchin — which was all he recognized. For this his
forecast was not wrong. What he overlooked — as businessmen did — was the drift of
908 BUSINESS CYCLES
It is, however, important to stress the common sense of the broad diag-
nosis thus invested with a spurious precision as to details. Realizing
from historical observation the extent of the revolution that had occurred
in the industrial structure and was in the act of upsetting its system of
values, shall we be surprised at the emergence of a situation in which
perhaps three-quarters of all businesses in the United States (including
farms) had to face the necessity of an adaptation that threatened them
with economic death? And is there really much to object to in the state-
ment that this situation was the fundamental fact about tfre world
crisis, compared with which all other factors, however important, were
after all but mitigating or accentuating accessories ?
Before proceeding to qualify and to elaborate, we will advert to a
consequence which follows if that diagnosis can be established. Capital-
ism and its civilization may be decaying, shading off into something else,
or tottering toward a violent death. The writer personally thinks they
are. But the world crisis does not prove it and has, in fact, nothing to
do with it. It was not a symptom of a weakening or a failure of the sys-
tem. If anything, it was a proof of the vigor of capitalist evolution to
which it was — substantially — the temporary reaction. And in any case
it was — again, substantially — no novel occurrence, no unprecedented
catastrophe expressive of the emergence of new factors, but only a
recurrence of what at similar junctures had occurred before.
The first qualification is that, so far, the above argument covers only
the course of events down to the bottom of the depression, which, as we
shall see, occurred in the second half of 1932. Subsequent events raise
problems concerning recovery policies, which bar any statement at this
point.
Second, it should be recalled that we never either undertook to
explain or succeeded in explaining everything about any crisis or even
depression. There was, in particular, the important class of "under-
standable but nonessential incidents." In this instance they may be
exemplified by the activities of, say, Hatry or Kreuger and so on: even
German experience might have been somewhat different but for the
impulsive personality of the leading man of the Darmstadter und Nation-
albank. An example of more important elements of this class of phe-
nomena is the violence — though not simply the occurrence — of the boom
and the crash in the stock market of the United States, which forms no
part of the essentials of our process, yet powerfully influenced it. But
the boundaries of this class should not be extended too far. The building
things below that surface movement, i.e., the longer cycles and their phases. That failure
of predictions to come true may thus be used to illustrate the shortcomings of what we have
termed the single-cycle hypothesis. It also illustrates the necessity for forecasting of
understanding the logic of industrial evolution.
THE WORLD CRISIS AND AFTER 909
booms and their slackening from about 1928 on, for instance, or the bulk
of the difficulties in the agrarian sectors do not belong here but were
perfectly normal elements of that process.
The American debt situation and the American bank epidemics —
there were three of them — are in a class by themselves. Given the way
in which both firms and households had run into debt during the twenties,
it is clear that the accumulated load — in many cases, though not in all,
very sensitive to a fall in price level — was instrumental in precipitating
depression. In particular, it set into motion a vicious spiral within
which everybody's efforts to reduce that load for a time only availed to
increase it. There is thus no objection to the debt-deflation theory of the
American crisis, provided it does not mean more than this.1 The element
it stresses is part of the mechanism of any serious depression. But
increase of total indebtedness at the rate at which it had occurred in this
country is neither a normal element of the mechanism of Kondratieff
downgrades — repayment such as was effected by big concerns either from
profits or from the proceeds of bond and stock sales fits better into the
picture — nor in itself an "understandable" incident, like speculative
excesses and the debts induced by these. It must be attributed to the
humor of the times, to cheap money policies, and to the practices of
concerns eager to push their sales; and it enters the class of understand-
able incidents only if we include specifically American conditions among
our data. Similarly, bank failures are of course very regular (though still
not essential) occurrences in the course of any major crisis and invariably
an important cause of secondary phenomena, in particular, again, of
downward cumulative processes. Those epidemics cannot, however,
any more than the German difficulties in this field, be considered as
wholly explained by the ordinary mechanism of crises or by that mecha-
nism plus the fact of excessive indebtedness all round or even by all that
plus the stock exchange crash. The German breakdown is partly
attributable to very special circumstances of an extra-economic nature
without which nothing could have shaken such concerns as the Austrian
Kreditanstalt or the Dresdner Bank. The American epidemics become
fully understandable only if account be taken of the weaknesses peculiar
to the American banking structure, which made it succumb as no Euro-
pean system would have succumbed under similar circumstances — in
particular, the presence, fostered by legislation and public opinion, of a
large number of small and inefficient banks and the absence of anything
like the English tradition. Compared with this, the insolvency of foreign
1 Professor Irving Fisher, see, for example, Booms and Depressions, p. 85, does not
seem to claim substantially more than that. We may demur at his sweeping statement
that a fall in prices always impairs debtors' ability to pay. But "that over-indebtedness
and deflation were strong and indeed dominating factors" need not be denied.
910 BUSINESS CYCLES
debtors was, especially since it impinged primarily on strong concerns, of
minor importance.
Third, there were the external factors. In noticing them in Chap.
XIV, Sec. C, we found that their importance, so far as the causation
of the world crisis is concerned, may easily be overestimated. Some of the
most conspicuous political and social changes of the postwar time, such
as the Russian developments or the destruction of the Austro-Hungarian
Monarchy, no doubt interfered with the "normal" working of the
economic engine but have — excepting perhaps the breakdown of the
Viennese banking center — little if any bearing on the crisis. Shifts of,
and impediments to, international trade made many countries, and
especially England, less prosperous than they would otherwise have
been, and shaped, by thus weakening the organisms on which the crisis
impinged, very many details of the picture, but this is all. In particular,
it would be unwarranted to attribute any of the major features, let alone
causation, of the American depression to a "flood of imports." Imports,
on the contrary, fell rapidly at the critical time: they totaled $4.4 billions
in 1929, little over 3 in 1930, a little over 2 in 1931, and about 1.3 in 1932.
German exports were in 1929 and 1930 a nonnegligible factor. But they
and their possible effects were more a consequence than a cause. Nor
can the fiscal and social policies of England and Germany have directly
accounted for more than a weakened resistance to the impact of depres-
sion1— against which, moreover, must be set their restraining influence
during preceding prosperities.
We have seen, it is true, that the indirect influence of those policies,
certainly in the case of Germany and possibly also in the case of England,
on short capital movements had much to do especially with what in a
narrower and perhaps more proper sense should be called the crisis.
And the provisional solution that had been arrived at for the problem of
international payments was bound to break down in any major depression
and, before doing so, to accentuate its difficulties. We do not even
wish to palliate the influence of such monetary disorders as occurred
before the depression had the chance of tearing up the flimsy tissue
of pseudo gold currencies. The South American, especially the Argen-
tinian, disorders and the fall of the price of silver have undoubtedly
played some role.2 But all this looms so large because a depression
1 This is perfectly compatible with our estimate of their importance in general. The
above statement is intended only to guard against overestimation of their importance
in the causation of the great depression.
2 Gold movements are covered by short capital movements. In spite of devaluation,
the shrinkage of international trade, and repayments, the Bank for International Payments
estimated the total of international short credits at from 29 to 30 billion Swiss francs for
1934, roughly three times the prewar figure. But, as has been stated before, apart from
them gold cannot have played any major role.
THE WORLD CRISIS AND AFTER 911
occurred for other reasons. As a man may suffer from many ills and yet
for an indefinite time lead a vigorous life without being seriously incon-
venienced by them until, when his general vitality has ebbed away, those
ills or any one of them may suddenly acquire what to the specialists' eye
will seem paramount or even fatal importance; so the economic organism
always does bleed from many wounds which it bears lightly in three out
of the four cyclical phases, and which spell discomfort when one cycle,
distress when two, catastrophe when all the cycles are in the depression
phase. No doubt, external injuries were of unusually great importance in
this case,1 but explanation cannot be derived from them.
B. 1930. 1. The United States. — The businessmen and forecasters,
who in the fall of 1929 had made up their minds that nothing worse was
ahead than a "recession" not much more serious than that of, say, 1924,
cannot have been very disappointed by the general look of things during
the first half of 1930. At the beginning, stock prices rallied strongly,
security issues were large, signs of improvement showed in many spots,
money was easy. All that, except the easy money — prime commercial
paper rate fell from 4.89 per cent in January to £ per cent in August and
was 2.88 in December; for June it was 3.54 per cent — passed quickly,
it is true, and proved to have been a meaningless flurry, due perhaps to
the confidence that was so widely felt or to the organized effort to make a
stand that prompted some additional spending. But until about the end
of June business moved along on a but slowly falling level not much
below the figures of 1929 in practically all lines.2
The second half of the year presents a wholly different picture. What
was generally recognized as liquidation all round was the outstanding
feature. Rates of contraction quickened and comparison with 1929
figures became increasingly melancholy. As we have put it above
(Chap. XIV, last sentence of Sec. E) people felt that the ground under
their feet was giving way. There was, however, no panic or even alarm
until, late in the year, distress signals showed in the banking sphere, the
failure of the Bank of the United States (December)3 attracting particular
attention abroad. And not only totals for the year but also figures for
the end of the year were far from being catastrophic. Total corporate
issues were $5.473 billions — slightly above the level of 1926, though only
1 It should not be forgotten, however, that the crisis was nowhere else anything like
so severe as in the United States, the country most nearly free from injury by external
factors.
2 The business curve of the Harvard Barometer was, during those 6 months, almost
horizontal. The A curve ("speculation") had lost by May what ground it had gained
during the first quarter.
3 The writer has, however, not been able to convince himself that Europeans really
thought that this bank was something like the Bank of England or the Banque de France,
and that this belief materially influenced events.
912 BUSINESS CYCLES
about 55 per cent of the 1929 figure — or, exclusive of refunding issues,
4.5 billions, slightly below the level of 1926. Outside debits (133 cities)
with 137.5 billions were only a little below 1926 and about 14 per cent
below 1929, though contraction was severe from the beginning of July
to the beginning of September. Outside net demand deposits, as already
pointed out in the preceding chapter, neither were appreciably lower for
the year nor fell appreciably within the second half of the year — July,
8.117 billions; December, 7.911 billions. But All Other Loans had already
shrunk sharply from November 1929 to May 1930 and then continued to
decline. Some hoarding demand for money, increasing "money in circu-
lation" and total federal reserve credit outstanding, showed in December.
Number of banks suspended (1,345) was in fact more than double the
yearly average 1921-1929 (627), but 1,158 of them were nonmembers.1
Excepting July and August, there was net influx of gold in every month
which, together with the issue of additional national bank notes, brought
excess reserves of member banks to about $475 millions by the middle of
November — a position of the banking system as a whole that was tech-
nically anything but weak, although the value of collateral was already
seriously impaired.
There was nothing abnormal under the circumstances in the quicken-
ing decline of prices of finished products, which within the year fell by
about 10 per cent. The wholesale price index was pulled down by the
fall in the prices of semifinished goods and especially of raw materials — •
the December average of the latter was over 20 per cent below the Jan-
uary average. Money rate of wages fell considerably in agriculture, but
was (see below) substantially maintained in industry. In the fourth
quarter, however, reductions in individual industries were sufficiently
important to affect the general index, though they still left real hourly
earnings at a level higher than that of 1929. Weekly money earnings
decreased even for the first quarter, and weekly real earnings eventually
fell2 to 90 per cent above the prewar level. Wage bill continued its
descent from the peak of the third quarter of 1929. In New York State,
for instance, factory pay rolls declined by about 25 per cent from January
to December.
Total monetary labor income for the year as estimated in the study
made by the Department of Commerce in cooperation with the National
Bureau (S. Kuznets, National Bureau Bulletin, Jan. 26, 1934, p. 5) was,
however, only 7.9 per cent and total monetary income produced only
15.1 per cent below 1929. But net earnings of all corporations (excluding
1 The sum of deposits affected was, however, nearly 865 millions, more than four times
the 1921 to 1929 average.
2 For the cost of living index used in arriving at the above statements see International
Abstract of Economic Statistics, 1934, p. 205.
THE WORLD CRISIS AND AFTER 913
tax-exempt ones and life insurance companies) before deduction of
income taxes were by over 78 per cent below the 1929 figure ($1,960
millions), printing and publishing, foods, beverages, tobacco products,
chemicals, metal and metal products, paper, pulp and its products doing
comparatively well, textiles particularly badly.1 The most serious
features of the picture are displayed by the indices of industrial produc-
tion and of employment. Output of equipment and durable goods in
general, as reflected, for example, in steel ingot production, fell sharply
after May. Motor- vehicle factory sales were less than 2.8, as against
4.6 billions. The Harvard Society's index of volume of manufactures
declined by over 22, the Federal Reserve Board's index of employment
by over 16.2 per cent within the year,2 whereby the former almost
reached the trough figures of 192 13 and the latter slightly fell below them.
Mining did somewhat better than manufactures, and power produc-
tion was in the first six months above, and in the last six months not much
below, the 1929 level. Total construction as measured by contracts
(Dodge figures for 37 states) was over 20 per cent below 1929 and would
have been still lower if publicly and semipublicly financed building had
not kept up or even somewhat increased. But it is worth noting that
public utilities increased their expenditure on construction ($644 millions
as against 473 in 1929) and that residential building did not decline much.
It was, as we should expect, industrial and commercial building which
caused the fall in the total.
The brightest features of the picture are to be found in the sphere of
consumption, the surprisingly good showing of which has not always
received the attention it merits. Sales by department stores, while con-
sistently and increasingly below those of 1929, were still 102 per cent of
the 1923 to 1925 average and their December holiday trade in the New
York district — containing, however, one more selling day than in 1929 —
was only 4.5 per cent under that of the previous year. Considering the
fall in prices there obviously cannot have been decline in physical volume.
This is borne out by the behavior of carloadings of the l.c.l. category,
which in contrast to total carloadings declined but moderately. Con-
sumption of a number of articles, such as cigarettes and gasoline, electric
1 See S. Fabricant, National Bureau of Economic Research Bulletin, Apr. 11, 1985.
2 16.2 is the percentage by which the December figure was below the January figure,
which is what we mean by loss within the year. The production index, however, did not
decline monotonically, and the January figure, though the index is corrected for seasonal
variation, does not seem to represent the situation well, because it reflects restrictions
decided on at the end of 1929 and partly reversed at the very time when they affected the
index. So the average of the first 3 months was taken instead.
3 Wheat, cotton, cattle and hog receipts were but little below 1929: the agrarian sector
behaved according to form. The value of the wheat crop was only a little over 60 per cent,
that of the cotton crop less than 60 per cent, of the respective values in 1929.
914 BUSINESS CYCLES
current for domestic use, telephones, radio sets, and refrigerators,
increased or declined but insignificantly. To save space, reference is
made to the careful study of Mr. A. B. Tebbutt.1 The number of
business failures, consistently and increasingly above that of 1929, was
as yet far from alarming: about 2,000 a month, which is less than the
average number of failures from October 1921 to June 1922.
An attempt to answer the question how far these facts can be trusted
to reflect the working of our model (including "understandable non-
essentials") naturally divides up into two tasks. On the orte hand, we
have to ask whether expectations from our model are adequately borne
out. This is obviously the case. Even disregarding the exact coinci-
dence of dates with our experimental count, we readily see not only that
the history of that year as a whole is not badly rendered by the formula
"a recession sliding off into deep depression," but also that the economic
physiognomies of the two halves of the year, differing as they do char-
acteristically from each other, conform to our idea of such a situation in
every single symptom excepting the short-lived rally at the beginning.
The reader should have no difficulty in verifying this proposition as to
general contours — the behavior of price level, output, interest rates,
deposits, clearings, incomes, employment, and so on — and it will be
sufficient to draw attention to a few points.
That rally at the beginning was, as has been said and as will presently
be explained, due to organized effort, but too much importance should
not be attributed to that effort, because expectation would in any case
have been for fairly sustained business at that time. In particular, it is
perfectly in keeping with our schema that industrial money wages did
not during the first half of the year fall to any significant extent — the
reader will remember that there is no reason to expect wages to fall in
recession and that no inference about subsequent disturbance can be
drawn from their failure to do so. The ease in the money market was
also normal under the circumstances2 and a simple consequence of "busi-
ness deflating itself." It does not call for explanation by Federal
Reserve Board policy, although the latter no doubt contributed to it
(see below). Not less true to form were the effects of that ease: as nor-
mally happens in recession, it helped to keep up residential construction
1 Arthur R. Tebbutt, Behavior of Consumption in Business Depression, published as
No. 8 of the Harvard Business School Business Research Studies. For that and the fol-
lowing year this study gives a very instructive picture, although its details (such as the
behavior of candy and women's night attire) are sometimes not easy to understand. The
opportunity may be taken to refer also to other studies in the same series which, as for
instance those on the operating results of chain stores, present valuable information about
the later phases of the depression and about the subsequent recovery.
a We observe similar ease and a sharp fall in rates after the crash in 1873 and after that
of 1893.
THE WORLD CRISIS AND AFTER 915
and to induce certain types of investment, for example, by utilities, which,
however, acted also under another stimulus. The unequal — in the case
of some, especially of new, commodities negative — rates of restriction,
the severe slump in the second half of the year in the sales of "post-
ponables," and the behavior of consumption should be particularly
noticed.
Hence, although our methods do not enable us to formulate our
expectations numerically and although it is impossible to say whether
our series exactly behaved as they should have done according to our
theory, it is possible to say that they did so as far as we can make out.
Facts would not refute us even if we made bold to say that nothing but
our process had acted on the economic system, and they certainly bear
us out if we conclude that that process constituted the dominant factor,
while — witness all those undoubtedly very competent judges of business
situations who were unable to account for or to predict anything but a
brief recession — without that process the events of that year and of its
sequel would have to be explained in terms of such slogans as overproduc-
tion, oversaving, overinvestment, and so on, i.e., could not be explained
at all.
On the other hand, we must consider the possibility that other factors
influenced the course of things in such a way as to produce a spurious
verification. Attempts to influence the process have not been entirely
absent and must be taken into account even if, relying on previous argu-
ment, we discard the possibility that influences from abroad might have
significantly affected the American situation in other ways than through
the stock exchange.1 There were, first, measures in aid of agriculture
and an appeal to the household remedy of the party in power, the Hawley-
Smoot tariff. The effects of the former, though certainly not negligible,
failed even in the agrarian sector to modify conditions sufficiently to
make them diverge from expectation and cannot a fortiori have been
very important for the system as a whole. The effects of the latter,
partly counteracted by reprisals, may under the circumstances be equated
to zero.2
Second, there was the President's hortatory action, which was, by
staving off reductions in wages and by stimulating investment, "to make
1 That influence was exerted through stock exchange operations we, of course, do not
deny. But this is sufficiently taken account of by recognizing that boom and crash had
been more severe than they otherwise would have been.
2 The writer does not wish to deny that conditions in certain industries may have
been slightly steadied by the tariff or that a more thoroughgoing analysis might make out a
stronger case. But until proof to the contrary is forthcoming, he is inclined to rate both
the stabilizing and the dislocating effects of the Hawley-Smoot Act rather low, and net
effects still lower.
916 BUSINESS CYCLES
certain that the fundamental business of the country shall continue as
usual/*1 The interest for us of this and similar attempts consists mainly
in testing those theories that see in depressions nothing else but the result
of businessmen's moods, which, themselves ultimate data, shape business
situations by means of the cumulations and accelerations induced by
individual acts. In this case conditions were quite exceptionally favor-
able to success. The American business world was, as has been pointed
out, by no means overpessimistic at that time. It was in the habit of
looking for a lead to the heads of a relatively small number df concerns,
which, moreover, were big enough to be able to influence the situation
"mechanically" by their own action. Expansion would have been
eminently to their interest and certainly was what they actually wished to
see. Receiving themselves a lead from a political agent that was in no
way antipathetic to them and being, many of them, imbued with high-
wage theories, they, in fact, made an effort both by refraining from lower-
ing wages or from doing anything else that would have been suggestive
of depression, and by launching out into investment — the utilities and
railroads in particular (see Chart LIII) responded to the appeal and even
borrowed for the purpose, so that the "principle of acceleration" had
plenty to work with. Nor was the result simply nil. The case shows to
perfection what can and what cannot be achieved — and explained — in
this way. We have above noticed the "flurry" at the beginning of the
year, which may, though only in part, be attributed to that effort.
Third, we have seen that public expenditure was kept going and even
increased, especially by means of public construction. It has been
estimated2 that net federal income-generating expenditure for the year
was 251 and for the second half alone $450 millions. This is not negligible.
But it may well be doubted whether that part of the total which may be
reasonably assumed to have become fully effective during the year can
have influenced events materially. The Federal Reserve System, finally,
followed a policy of easy money — which policy, as we have seen, but
ratified the situation — thus giving business all the rein and all the encour-
agement it could possibly have wished for.3 As soon as it had become
1 Cf. President Hoover's message to Congress, December 1929. For similar utterances
by various types of authorities see Mr. Walter Lippmann's quotations in The United States
in World Affairs, 1931.
* See Professor A. D. Gayer, What is Ahead?, The New Republic, Feb. 2, 1938, p. 391.
The federal deficit for 1930-1931 was over 900 millions. This was mainly — to almost
three-quarters — due to a fall in revenue, while in England and Germany revenue increased,
in England's case by 43 million pounds; in Germany's case insignificantly. Taxation may
by mobilizing idle deposits be income-generating, even if it entail no borrowing from banks.
'The question whether the "dear money policy" of 1928 and 1929 can, partly or
wholly, be held responsible for the events in 1930 has been discussed in the preceding
chapter, Sec. F; the questions of principle involved, in Chap. XIII.
THE WORLD CRISIS AND AFTER 917
obvious that business was in for a "recession" and before any vicious
spiral had developed, the reserve system resorted to what the public and
many economists had by then come to believe in as the remedy, open-
market purchases on a large scale. From October 1929 to December
1930 it bought government securities to the amount of 560 millions or,
more precisely, its holdings of government securities rose from 136
millions on Oct. 23, 1929, to 533 millions on Dec., 18, 1929, then fell to
477 in January 1930 to rise again to 602 by the end of August. After
that date, purchases were reduced to insignificant amounts. For this
there were two excellent reasons. First, those purchases which it was
thought had been so effective in stimulating activity in 1924 and 1927,
at this time did not seem to have any effect at all.1 Members reacted
primarily, although they also increased their investments, by paying
off rediscounts, and were obviously much more troubled about finding
customers for their funds than about finding funds for their customers:2
to anyone at all open to argument and evidence on the subject, any
further steps in that direction would have seemed altogether futile.
Later on, in the second place, the gold influx provided another reason for
discontinuing that policy.3 We may, hence, conclude that the behavior
of the reserve system, while it favored cheap money and expansion and
certainly exerted no depressing effects, yet was no major factor in shaping
the business situation — our process seems, as far as that goes, in fact to
have worked all but undisturbed.
2. England. — The outstanding fact about the English depression is its
mildness, which makes it doubtful whether that term is applicable at all.
This statement may sound strange to many readers who are, in connec-
tion with the world crisis, impatient of anything but superlatives. It
could, however, be reduced to the obvious if we consider that the behavior
of some of the worst symptoms is amply accounted for by the fall in
exports and the impact of foreign insolvencies — phenomena that were
1 This will not astonish us. But it is not superfluous to weigh the implications of that
fact, which should according to ordinary rules of inference suffice to cast doubt on the two
previous instances.
2 Rediscounts had been at about 1 billion in the summer of 1929 and were down to
about 200 millions by July 1930. Investments of reporting member banks rose from
5,486 millions on Aug. 28, 1929, to 6,829 millions on Aug. 27, 1930. The reader should
observe how very "regular" all this was from our standpoint.
8 For the year net gold imports were about $280 million. Net release from earmark was
minus 4.4; total increase in gold stock 305.4 millions. An eminent economist, who is
more enthusiastic about open-market operations than the writer feels able to be, attributes
their failure to take effect to inadequate dosiqg and to the policy's being abandoned "in
the face of" the gold imports from October on. This implies that he looks upon gold
imports not only as factors of decrease of Federal reserve credit outstanding in the technical
sense, but also as a reason for and not against open-market purchases. This the present
writer entirely fails to understand.
918 BUSINESS CYCLES
to be sure largely due to cyclical factors (and their "secular" net results)
but must be classed as external from the standpoint of the English
organism.1 But, even independently of that, the depression was in
important respects much milder than in this country. Nor shall we
wonder at it, for if the evolutionary process is the fundamental "cause"
of prosperity and depression, relative weakness of its positive phases will
be accompanied by relative mildness of negative phases.
In 1930 the London and Cambridge Economic Service annual index
of total production fell to a little below the figure of 1928 (1928, 108.7;
1929, 115.8; 1930, 106.5). Some groups, however, displayed increase
(leather and India rubber trades) or insignificant decrease (food, drink,
tobacco; nonferrous metals).2 Within the steel group, the number of
motor vehicles produced declined very little.3 Analysis of the building
index, which fell more than any other component of the total, reverses the
inference to be drawn from it: the fall in houses completed during the
year is entirely due to the cessation of construction under the Chamber-
lain Act and, to a lesser extent, to temporary reduction of building under
the Wheatley Act, while the joint effect of this was to more than 50 per
cent made up for by privately financed construction. This increase in
unsubsidized housebuilding, which gathered force instead of slackening in
the second half of the year, suffices by itself to negative the idea of a
depression of unprecedented severity. With this the high and rising
unemployment figures seem difficult to reconcile. But the fall in exports
(see below) partly resolves this apparent contradiction.4
Wholesale prices and cost of living fell steadily, as we should expect —
the fall in the former, the reader should recall, meant partly, though not
wholly, a gift presented to England by the countries producing raw mate-
rials— and short money rates continued during the first 5 months their
1 It is disheartening to observe that difference of opinion is possible about what seems
to the writer so obvious a point. Mr. Hawtrey (Trade Depression and the Way Out, p. 27)
for instance, state that in the latter part of 1929 England was already a "deadly centre of
contagion" from which the "blight of pessimism" was ready to spread. This is on a par
with his statement (ibid., p. 26) that depression had prevailed in England ever since 1920
(sic). In part such differences are presumably due to different use of terms. But the
present writer feels unable to understand those statements — and many others — in what-
ever sense he can conceive them to have been meant.
2 The index includes agricultural output, which also declined, though not quite so much
as the total index.
8 So did the shipbuilding index of the Service. It is, however, important to note that
tonnage commenced declined drastically in the second half and especially toward the
end of the year. Coal also declined (see below), but less than steel. Conditions in the
textile industry were unsatisfactory.
4 From January to December 1930 the total number of unemployed insured persons
(Great Britain and Northern Ireland) increased by nearly 700,000. The average figure for
the year was 16.1 per cent. Part of this, however, was an effect of previous rationalization.
THE WORLD CRISIS AND AFTER 919
descent from the peak in the autumn of 1929. After that they fluctuated
at a low level, the three months' rate, for example, around 2.3 per cent,
which was about 1 per cent below bank rate.1 Deposits in the nine clearing
banks moved above 1929 figures in the second half of the year, as did
investments and discounts. Town and metropolitan clearings were but
moderately smaller, but provincial clearings by over 15 per cent. New
capital issues for Great Britain were less by 20 per cent. Prices of fixed-
interest securities rose, and prices of industrial stocks fell in the way that
is normal under the circumstances. Professor Bowley's index of average
weekly wages, which had declined by one point and a half in 1927, by
two points in 1928, and by one in 1929, fell by another point and a half,
and thus only continued in 1930, without acceleration, a tendency it
had displayed before, while real wage rates, of course, rose considerably
within the year.
The reader should judge for himself how far this pattern accords with
our model. In doing so he should observe that in order to account for
the very great difference between it and the American pattern we need
not fall back upon anything external to that model. Nor could we do so
if we wished, for although there is no doubt, of course, that the English
organism was much more exposed to external influences (in our sense)
than the American, nothing but depressing effects could have emanated
from them. The most important has already been mentioned. Money
value of total exports fell off by approximately one-fifth (1929, 720
million pounds; 1930, 571; peak 1924, 801), that of all manufactures by
about one-fourth, the greater part of this loss being concentrated on the
second half of the year. Income from foreign investments also declined.
Domestic policy may for the purpose in hand be described as neutral.
The slum clearance and the coal acts exerted, the first no effect, the
second little effect during the year.2 The budget for 1930-1931— it
was Mr. Snowden's second budget, though the first for which he took
full responsibility — provided for a considerable increase in ordinary
expenditure, most of which was, however, due to the "derating plan"
of 1929-1930, i.e., the transfer of certain local burdens to the national
budget, which only then took full effect. Income tax was increased by
sixpence, allowances for smaller incomes being so adjusted that the
increase was effective for only about one quarter of all income tax payers,
1 Bank rate was reduced to 3^ per cent on Mar. 20, 1930. The maximum of 6^> per
cent occurred, it will be remembered, Sept. 26, 1929-Oct. 31, 1929. At the beginning of
1930 it was 5, from Feb. 6, 4^; and from Mar. 6, it was 4 per cent. Again, as in the case
of the United States, it is assumed that the question of the influence of 1929 rates has been
disposed of in the preceding chapter, Sec. F. The gold situation, which did not grow worse
until November, will be more conveniently dealt with later on.
2 Average net selling value of coal at pit head and wages per man shift were, however,
both slightly higher at the end of December 1930 than at the end of December 1929.
920 BUSINESS CYCLES
and the surtax and inheritance tax were raised, the additional burden
being considerable only in the higher brackets. Without receding from
his general opinion about the effects of fiscal policies of this type, the
writer believes that the increments decided on cannot have accentuated
the depression materially and that, on the contrary, the budget may have
strengthened confidence in the seriousness of labor finance.1 All this
was not "pump priming." But neither was it "deflation."
3. Germany. — Though German developments during that year,
unlike American, but like English developments, could n&t be com-
pletely understood without appeal to external factors, fundamental
contours nevertheless answer to our schema. To factors external
to the economic process we attribute, as has been explained in Chap.
XIV, Sec. E, the early emergence of difficulties right after the strong
prosperity of 1927. German industrial production and employment
fell already in 1928 — more precisely from the last quarter of 1927 — and
the former but weakly indicates the bulge in 1929, which is so marked in
English and American graphs, while unemployment in that year was
getting nearly as serious as it had been in 1926. Prices of shares had
declined — this, of course, was perfectly normal, much more so than the
American boom — ever since April 1927, and so had issues of fixed interest
securities — this was contrary to what we should expect to happen within
our process in a Juglar recession, though readily explainable under the
circumstances — and government as well as municipal finances had begun
to be a source of anxiety in 1928. It is thus not surprising that people
were talking about a "crisis" from the beginning of 1929, general uneasi-
ness being intensified by failures of the type that reveals unsound financial
practice (Frankfort Insurance Co., August 1929, the failure of which had
nothing to do with insurance but all the more with promotion — the
German Hatry case), by the acceptance of the Young plan,2 by the
sudden withdrawal of foreign balances in April and May 1929, by com-
munist unrest issuing in an outbreak in the latter month, by the swelling
1 The deficit eventually amounted, as per accounts, to 84 millions, but disappears
when properly corrected for items which should not be counted in figuring out a deficit.
On Consolidated Revenue and Expenditure Account (cf. Colin Clark, op. cit., p. 140)
of Government, Local Authorities, and Social Insurance, there was an excess of Income
(1,022 million pounds) over Expenditure (982 million pounds). Unless we assume that
all of that billion was paid out of existing and active deposits, we are driven to the conclu-
sion that public expenditure added, though only a little, to the sum total of income, i.e.*
that there was a small amount of "income generation" through public expenditure.
8 Whatever intentions and the merits of the case may have been, the Young plan sapped
the political strength of the government and the parties responsible for its acceptance
much more than they themselves at first realized. Though it did not immediately create
any difficulties, and though it fully provided for all difficulties which were thought likely
to arise, it contributed — though under rather than on the surface — to that state of the
public mind and is thus relevant to our subject.
THE WORLD CRISIS AND AFTER 921
f social burdens, by the inability or unwillingness of government to
nth the labor situation, and by a spreading sense of the brittleness
political structure of the time.
hile objective facts do not quite bear out that pessimistic attitude
129, the situation deteriorated in 1930 more rapidly in Germany
in England or the United States. It is true that public spending
»elow) kept up consumption and that sales of cooperatives — even
il retail trade in such articles as furniture — fell off but little. It is
rue that hourly wage rates did not fall till the last months of 1930,
i the contrary slightly rose until then, keeping on a level of nearly
• cent above 1925 and, for the year, also above 1929. But this does
tiaracterize the situation, because they were so kept by political,
and administrative pressure, particularly by the practice developed
e official arbitrators (Schlichter), and because, these official rates
trade union minima, the rates actually paid may have fallen
ut affecting wage statistics, provided they still remained above
minima.1 According to an investigation of the Verband Deutscher
lindustrieller for the Berlin district, the latter, however, does not
to have been the case, at least to an appreciable extent, until the
in of 1930. No doubt, of course, can be entertained about the
,se of real wages, the Reich's cost of living index (1913-1914 = 100)
$ at an accelerating rate from 151.6 in January to 141.6 in December,
number of symptoms behaved in a genuinely normal way. Bond
rose until July — the London quotation of the Dawes loan began to
May — when the political situation provided a reason for a down-
money rates fell till about August, the bank rate staying at 4 per
the postinflation minimum, from June to August; Reichsbank
igs and debits to postal accounts receded moderately; advances in
it account (Schuldner in laufender Rechnung; Grossbanken only)
,sed and so did the amount of bills held by banks, though under
tions of monetary ease the Reichsbank's portfolio decreased sharply
I the first three quarters of the year.
le gold withdrawals from middle of September to middle of October,
nting to about 1 billion marks, were met and the reserve was
dshed to the extent of half of that sum from the proceeds of a foreign
^ency loan negotiated by the federal government. They were
Y political in nature.2 The increase in bank rate which they induced
so, under conditions of severe unemployment, actual payments below those minima
>ssibly have been connived at by trade unions. The "rigidity of wages" thus may
ious in part.
is just possible that small withdrawals would have been made in any case, owing to
ceding fall in German money rates. But since in other countries rates had fallen as
>r more, this is not likely. Nor was it likely, circumstances being what they were,
922 BUSINESS CYCLES
led to a stiffening of rates all round, but the situation being what it was,
an increase of lj^ per cent (prime bankers' acceptances) can hardly have
been a major consideration in short-term operations, while nobody was
ready to embark upon long-time investment in any case.
Total industrial production shrank, relatively, as much or more than
it did in this country, and it did so from the very beginning of the year.
The January figures for crude steel and products of rolling mills were
down by about 13 and 10 per cent as compared with the "figures for
January 1929, which were already low.1 By December the (imperfect)
index of production landed at 72.5 per cent of 1928. The total number of
unemployed was about 3 millions during the first 10 months and rose to
nearly 4.4 in December. There cannot be any doubt that the extra-
economic factors alluded to had much to do with this. Superimposing
themselves on what would in any case have been a depression, they
intensified it greatly, as some of them had previously damped prosperi-
ties. The writer does not know of any facts not covered by this diagnosis.
In particular, it is (when we recall what those factors were) reasonable to
say that they could not have produced a slump of that magnitude by
themselves, i.e., if the organism had been in one of its positive phases.
Government was influenced by two opposite considerations or, rather,
was under pressure from two different camps. Many interests, especially
the agrarian interest, the condition of which grew steadily worse, and the
labor interest, which anticipated a breakdown of the unemployment
insurance and further increase in unemployment, needed and pressed for
help. There was also increasing demand for pump priming. Other
interests pointed and pressed in the direction of an attempt to use the
depression in order to normalize general conditions, i.e., to break with the
spending habits of public bodies, to retrace steps in social legislation, to
reduce money wage rates, to buttress the currency and its purchasing
power, and so on. Such views were not, however, held by interested
parties only. Many people even believed not only that such a policy
would provide sound foundations for future progress, but also that it
would have remedial effects at the moment. "In the years from 1927
to 1929 we have been wandering in an economic, financial, and socio-
political labyrinth (or maze; the German word was Irrgarten) . And
now we must return to sober reality" — so the federal minister of labor,
that an increase in bank rate of one per cent would exert any protecting or attracting
influence. As far as that goes, the increase was merely ceremonial. And since there was
no excessive activity to restrain, it is difficult to see what purpose it was to serve. Econo-
mists imbued with a mystical belief in the efficacy of interest rates will no doubt explain
all that was to follow by it. This cannot be helped.
1 Coal output was, however, higher for January and February, output of coke for
January.
THE WORLD CRISIS AND AFTER 923
Stegerwald, himself a prominent labor leader,1 declared as late as Decem-
ber 1931. Whatever the merits or demerits of such a policy, which
has been and is so unintelligently recommended by some and so unin-
telligently rejected by others, it is important for our purpose to ask how
far it was proceeded with during that critical year. In doing so we must
realize that those two policies were not so mutually exclusive as they might
seem. Certain steps on the lines of the first were hardly avoidable
precisely in order to get maneuvering space for the second. Some steps
on the lines of the second even an adherent of the first would have had to
take if he was to avoid chaos in a country which had had so recent an
experience of what inflation is. A horseman who alternately collects
and gives to his mount does not thereby contradict himself or prove his
lack of purpose.
The government spent freely. The federal deficit of over 1 billion
marks (1930-1931) following upon a deficit of 712 millions (1929-1930)
was relatively more income generating than was the deficit of the United
States. In taking measures in aid of agriculture, in providing means for
those unemployed whom the statutory or financial limits of unemploy-
ment insurance failed to include, in doing some pump priming and
straining its credit (the slogans were cranking, Ankurbelung, and bridging
credits, Ueberbruckungskredite), the government certainly discovered and
acted upon the theory of the unbalanced budget. By the (amendment
of the) bank act of Mar. 13, 1930, long-term bonds of the Reich, of the
states, and of the municipalities became eligible as collateral for Reichs-
bank credits, and toward the end of the year treasury bills were freely
discounted by the Reichsbank. This is not exactly " deflation," whatever
that term may mean. Two moves on the other line were obviously
intended to serve as safeguards in the long run. First, the shadow of a
curb was provided for future excesses of municipal finance by certain
regulations intended to make it more difficult for municipalities to run
into debt (Dec. 3, 1930). Second, the Reichsbank nailed its colors to the
mast by putting into operation Clause 31 of the bank act of 1924, which
bound it to do what, so far, it had done without obligation, viz., to
redeem its notes in gold or exchange (Apr. 19). Both measures might
have become effective in the future, but certainly cannot have had effects
in 1930.
The only other point that matters here was the government's move to
reduce rigid prices (end of August), which went parallel to its efforts to
keep up agrarian prices. The index of prices at wholesale (Institut fur
1 He was not a socialist. But socialist labor leaders did not object as strongly as they
assuredly would have done if they had really thought him wrong. The writer's impression
is that they too were of the opinion that things had gone too far and rather felt relieved
when somebody else shouldered the unpopularity of leading back to the highway.
924 BUSINESS CYCLES
Konjunkturforschung: 1913 = 100) for 1930 was 114, as against 131 for
1929, and fell rapidly within the year. But the government and the
public were as much exercised about "administered" prices as the
American government and public are now. Moreover, the argument
presumably suggested itself that it would be easier to reduce wage rates
if the government aimed also at other points in the price structure.
Accordingly, it embarked upon a campaign for the reduction of prices and
costs, while the Reichsbank busied itself until October in bringing about a
general reduction in interest rates, particularly in East Prussia.1 We are
not concerned with the logic of all this. Since it worked with and not
against the stream, it probably effected little beyond somewhat quicken-
ing adaptation in sticky spots. A number of rigid raw-material prices,
which had not moved at all during 1929 and the first months of 1930,
fell by a little over 10 per cent toward the end of the year. The weighted
average of trade union wage rates declined by 5.8 per cent in the first
4 months of 1931, while the index of production increased by 9.1 per cent
in the first quarter, and employment by 5 per cent during April and May
1931, both indices being cleared of seasonals. Presence of a causal rela-
tion was claimed, possibly not quite without foundation.2
C. 1931 and 1932. — While it was necessary to have the facts of 1930
firmly planted in the reader's mind, it seems possible to confine discussion
of the two years which span the real "catastrophe" — of the capitalist
system or of an individual system of economic values — to a number of
comments that can easily be worked up into a connected survey.3
1. Physical Production. — The fundamental point to emphasize is
again that there is nothing in the fact that severely depressive symptoms
1 Annual report to the shareholders' meeting, Apr. 29, 1931.
2 So, for example, in Lohnpolitische Kurvenbilder zur Krisenlage, Verein Deutscher
Maschinenbau-Anstalten 1931 (mim.). Even the "possibility" that that argument be
"not quite" without foundation really goes much further than most modern economists
would care to go. The present writer would not go further than that. But that possibility
he is prepared to prove. That pamphlet does not fail to notice the lag between increase of
output and increase of employment, which it explains partly on statistical grounds and
partly on the ground that in such situations the labor employed is underutilized.
8 There is some danger in thus leaving room for uncertainty of interpretation on many
points which could be removed only by thorough discussion of every detail. But we have
no choice. Some general impressions can be gleaned from referring to our charts. The
current reports in the Review of Economic Statistics (W. L. Crum) and the London and
Cambridge Service would prove helpful. Beyond this, the simplest way of testing and
supplementing the statements of our text is perusal of the World Economic Surveys by
Professor J. B. Condliffe (League of Nations, Economic Intelligence Service). Professor
Bertil Ohlin's Course and Phases of the World Economic Depression will prove a valuable
guide up to the summer of 1931. Comparison between those interpretations and the one
offered here is invited. Mention should be made, also, of Professor Lionel Robbins's
analysis in The Great Depression.
THE WORLD CRISIS AND AFTER 925
continued to dominate the picture in all three countries to require addi-
tional explanation from our standpoint — nothing has to be explained
away, and that fact fully conforms to expectation from our model. As
stated in Sec. A, that picture reflects ever since, roughly speaking, the
middle of 1930 an element not in evidence before — the "vicious spiral"
which for nearly two years would by itself suffice to describe the surface of
what was going on. But we know that the process thus designated is to
a greater or lesser extent a feature of any depression. It is, in fact, largely
responsible for turning the "normal liquidation" of recession into the
"abnormal liquidation" of depression (Chap. IV). At that stage> the
phenomena expressed by the "principle of cumulation or acceleration"
and by the "debt-deflation" theory become part of the cyclical mecha-
nism, a particularly important part when all three cycles are in the phases
most favorable to them.1 But since we also know that the working of
the spiral is erratic ("internally irregular") and extremely sensitive to
incidents, accidents, and external factors, and that the troughs it creates
are intrinsically unreliable, we must, especially in the face of the various
recovery policies that were taking shape in those two years, recognize the
limitations of any attempt at verification beyond that fundamental
fact. Barring crises and panics, which may occur at any time and the
occurrence and effects of which have simply to be registered, we expect a
depression to display shrinkage at decreasing rates until it shades off into
recovery, in the case of the Kondratieff on a very broad bottom and in the
case of the Juglar on a bottom of about a year, while the Kitchin phases
should show in what statisticians call surface movements in series cor-
rected for seasonals. Even without attaching weight to the dates yielded
by our experimental count (see above, Sec. A) we will note that since a
Juglar depression must contain positive Kitchin phases — and, if we trust
our schema to that extent, end up with a Kitchin prosperity — it should
display an observable, though possibly weak and short, upturn at the
bottom, which again is nothing else but our way of expressing a familiar
fact of business experience.
1 Hence, it would not constitute valid objection to say that the events of these years do
not require explanation by the theory of innovation but are adequately explained by the
vicious spiral or by the principle of acceleration. The one and the others do not stand on
the same plane of argument and cannot be pitted against each other. Believing he has
made this abundantly clear throughout this book, the writer wishes to add only two minor
points. First, the word acceleration seems to suggest increasing rates of change; this
would, however, even if the word were to be taken at face value, not contradict the expecta-
tion to be presently met with in the text, because that word refers only to a component
and the expectation to a resultant. Second, it might be held that, the facts covered by the
term Vicious Spiral being the outstanding feature of those years, we may stop at that and
discuss the theoretical and practical problems of the situation without going beyond those
facts. This is not so. It is not indifferent, for either diagnosis or therapy, what starts the
spiral.
926 BUSINESS CYCLES
What are the facts? According to a very general opinion with which
we agree in this case, a preliminary answer may be derived from the
behavior of industrial production: nothing else qualifies equally well for
the role of an indicator of the objective state of the system in the later
part of depression and during recovery. We will look first for the lower
turning point, or rather, remembering our view about troughs, for the
lowest segment of our — or almost any1 — graphs. In almost all countries,
however different their structures and general conditions,2 particularly in
Austria, Belgium, France, Germany, Hungary, Italy, Poland, and
Sweden, it occurs in (the middle of) 1932. In the Canadian index the
trough comes in February 1933. Japan was an exception all along, in
fact, the standard instance for advocates of inflationary policies, who
have only to be reminded that the case is a special one in that this policy
found all the conditions ready at hand for a rapid industrialization of the
country. In the English case there is some doubt, which will, however,
be presently removed;3 but our chart displays a well-marked trough in the
summer of 1932. So it does for this country. The relapse in the spring
of 1933, which carried the American index back to about its previous low,
while not in itself astonishing — that after a depression of such severity
relapses should occur in the first stages of recovery is perfectly under-
standable on common-sense grounds, even without appeal to the course
of Kitchin phases — was serious enough to suggest the presence of dis-
turbing factors. But since the advent of a new administration pledged
to pursue an active policy and the third epidemic among banks — which
will be dealt with below — readily supply the explanation, it seems reason-
1 Graphs displaying the behavior of indices of production in different countries are, in
spite of the problems that arise on the score of comparability, among the most common
tools of scientific and popular analysis of the crisis. Specific mention should, however,
be made of the work of the National Bureau — Professor Mills in the Bulletin for Feb. 20,
1933, and Mr. Bliss in the Bulletins for June 26, 1934, and Nov. 15, 1935— and of Miss D.
Westcott, Review of Economic Statistics, Dec. 15, 1934. Also see again N. J. Wall, Monthly
Index of World Industrial Production, 1920-1935, Bureau of Agricultural Economics, 1936.
2 Unfortunately, it is not only the economic structures and conditions but also the
indices which are different to the point of incomparability. It is believed, however, that the
validity of our argument is not substantially impaired thereby. The shapes of the bottom
parts of the curves are not alike and raise various questions of detail, which will, however,
be dealt with only for the United Kingdom and the United States. If we can trust the
index or the available indications, Rumania and Spain did not conform. But the course
of things in Spain was clearly politically conditioned (proclamation of republic, April
1931). In Tchechoslovakia the bottom comprises 1932 and 1933, but the steepest descent
was in the last quarter of 1931 and the first quarter of 1932, after which the index crawled
along level until late 1933.
8 In the London and Cambridge Service annual index (including agriculture and build-
ing) the low occurs for 1931, but 1932 displays the annual minima in iron and steel, non-
ferrous metals, food, drink, tobacco, and the leather group. In mining 1933, in textiles
1930 is the minimum year.
THE WORLD CRISIS AND AFTER 927
able to accept the 1932 trough as the "true" one, although according to
some indices — not, however, that of the Federal Reserve Board — the
trough in March 1933 is somewhat deeper still. For both manufactures
and minerals, also for steel, lumber, petroleum refining, coke (by-product),
foods (the descent of which, however, was hardly perceptible) textiles,
automobiles, and construction (values) taken separately, decline came to
a stop by roughly the middle of 1932. Only for cement, rubber products,
and power production,1 possibly also tobacco, the halt does not come
before the spring of 1933, while leather and leather products turned up by
the end of 1931. The behavior of carloadings and imports (current
values) bear out the general picture and so did business failures, which
fell sharply after reaching their peak in the summer of 1932. The
Federal Reserve Board's index of department-store sales declined until
the spring of 1933, but this was due to the continued fall in prices.
Thus it seems that, so far as physical production is concerned, the
location of the bottom of the depression answers well to our idea as to
where it should have occurred. Turning, in the second place, to the
descent to that bottom, we find2 that production in the United Kingdom
suffered its most severe contraction in 1930, and that the rate of shrinkage
was — even disregarding the little hump between the autumn of 1931 and
the autumn of 1932 — smaller for 1931 and still smaller in 1932, the move-
ment retaining its comparative mildness throughout. A smooth curve
fitted to the German figures from the middle of 1930 to the end of 1932
would display a monotonically decreasing rate of fall, shading off into
increase. American industrial production declined for 1932 as a whole
very much less than it did in 1931 and, owing to the upturn at the end,
the average rate of change within the year would also be smaller. But
the contraction at the beginning of 1932 was (in percentage rate) the most
serious of all, and whether the decline of 1931 was milder than that in the
second half of 1930 is a matter of index construction.3 These irregular-
1 Electrical power production did, however, strike a plateau during the two middle
quarters of 1982, before it took another dive. As has been mentioned, sales for domestic
use continued to increase even in 1932.
2 See, for example, D. Westcott, op. cit., p. 256, or our own Chart XLII. For Germany
Miss Westcott's curve is better than ours. The total decline from the 1929 peak to the
trough was, of course, much greater in the United States than in most countries. This is
certain, in spite of all the doubts about comparability. Mr. Bliss (07?. cit., p. 2) estimates
it at 53 per cent, while for Germany the corresponding figure is 43 and for the United
Kingdom (on quarterly data) 22 per cent. We will add the percentage fall of crude steel
production for the first 9 months of 1932 as compared with 1929: Germany, 66.4; Great
Britain, 46.3; United States, 76.5.
3 According to the index chosen for our chart, decline for the first half of 1930 was no
less precipitous. But this is clearly out of keeping with the general aspect of business at
that time.
928 BUSINESS CYCLES
ities may, however, be partly accounted for by the specifically American
difficulties in the sphere of credit and banking.
We cannot stay to note various other points of interest.1 One fact
only which sheds much light on the nature of the process cannot be passed
by, viz.y that the depression acted as an efficiency expert. This holds
true of all three (and other) countries, but it is particularly in evidence in
the United States. Output per man-hour, which may be roughly said to
have increased by 22 per cent from 1923 to 1929, surpassed in 1932 the
1929 figure by an amount that will greatly vary according to the indices
entering into the computation, but may plausibly be put in the neighbor-
hood of 20 per cent.2 This was, of course, not only due to rationalization
under duress in the concerns that continued operations, but also to the
shutting down or permanent elimination of others, which may as a
broad rule — though, as we know, by no means always — be assumed to
have been less efficient. The second component, however, enters our
model no less than the first. More rigorous selection of the work-
men employed or to be employed may also have had some, the mere
underutilization of fixed factors cannot have had significant effect — the
latter factor is likely, under the conditions of modern industry, to have
worked the other way, at least in many cases. Reference to Chart
XLIII is all that is necessary as regards the behavior of group indices.
It was as we should expect, foods and perishable commodities in general
keeping up comparatively well throughout the bottom of the depression.3
For most countries total industrial as well as manufacturing produc-
tion experienced a beginning of recovery, as small as we should expect,
1 See, for example, Charles A. Bliss, op. cit., p. 10, for the wide dispersion of percentage
changes 1929 to 1932 in individual items. Note the increase in refrigerators and rayon and
the particularly sharp decrease in railroad cars, locomotives, pig iron and iron ore.
2 Professor Mills (op. cit., p. 4) arrives at 12 per cent, Mr. Bliss (op. cit.t p. 6) at 21 per
cent (manufacturing only), which in the second publication quoted before he reduces to
18 per cent. See the comments of both. It is in accordance with our view of the processes
of negative phases that, as both authors state, this increase was not due to any significant
extent to "revolutionary" innovations. But if both point to the decline in the production
of machinery between 1931 and 1932 in support of the contention that installation of new
equipment was not a dominant factor in that increase in output per man-hour, we must not
forget that installation of new equipment exerts its influence with a lag. This is so even
within the individual concern, but still more in industry taken as a whole, because the
process of adaptation and elimination of less mechanized or otherwise less efficient concerns,
which is the backbone of depression, is a slow process.
8 For details on the behavior of household consumption, see again A. R. Tebbutt, op. cit.
In particular, it is interesting to note that deflated sales of furniture at department stores
in 1931 were below 1929 by no more than about 12 per cent. Toilet articles and drug
sundries even kept their dollar volume substantially intact. Cigarette consumption did
not decline substantially before 1932, when it still kept about the 1928 level. Gasoline
consumption rose in 1931 and fell in 1932 by only 10 per cent.
THE WORLD CRISIS AND AFTER 929
in the second half of 1932,1 which slackened toward the end of the year,
when, as we would say, the last Kitchin of the Juglar turned into reces-
sion. Such a relapse at that juncture is in itself no problem. The
majority of countries for which indices are available display it, but not
Germany — nor France, where, however, industrial production embarks
upon a decline in the middle of 1933 — whose production moves up steadily
and strongly from the third quarter of 1932 to the middle of 1934. In
the United States production of manufactures and mining as measured
by the seasonally adjusted index of the Federal Reserve Board increased
substantially during August and September — from the July low of 58 per
cent of the 1923 to 1925 average to 66 per cent — and went on at that level
in October; then, after a slight setback, regained it in December. It is
important to note that it was the activity of cotton, woolen, silk, rayon
mills, shoe factories, and the like which was responsible for the increase in
August, while automobile production still declined and steel and lumber
industries did not show even seasonal advance. In September, though
the steel industry expanded a little — it then reached 20 per cent of capac-
ity during the first three weeks of October — the same general features
characterize the situation. In October, automobile production increased,
but there was still little activity in investment industries. This is in
accordance with the usual pattern. We know that in a four-phase cycle
recovery from the lower turning point does not typically start either by
innovations or, more generally, investment, or by firms' borrowing but
exclusively by moves, from an indefinite number of points and within the
existing framework of plant, equipment, and balances, toward a neighbor-
hood of equilibrium.
It remains to account for the upturn which occurred in the industrial
production of this country and of Germany, but not in that of the United
Kingdom, at the beginning of 1931. In this country more than in others
stock prices, pay rolls, loans, velocity of deposits, and department-store
sales increased along with production, while employment at least ceased
to fall for over 4 months. Readers may well feel that we need not bother
about this intermezzo, for it is easy to understand that so long a decline
should be interrupted by temporary rallies. Optimistic anticipations of
people about "recovery being around the corner" may also count for
something in the explanation of a phenomenon of this order of magnitude.
But there may be more to it. No recovery was due as yet, but a pro-
gressive abatement of depressive symptoms was. The elimination, by
xThe indices of world production of the League of Nations (excluding Russia) and
of the Bureau of Agricultural Economics (N. J. Wall, op. cit., p. 2) turn up exactly in the
middle of the year. So do the Austrian, Belgian, French, and Swedish indices. In Italy
and Hungary the upturns occur a little earlier. The Polish index rises from the beginning
of the year, but falls to a new low in December.
930 BUSINESS CYCLES
the processes of depression, of untenable situations and of obstacles that
give way only under duress, was partly accomplished. Some adjustments
in the cost structures of the kind noticed in our description of German
events of 1930 may have given an impulse to individual firms and indus-
tries. If so, subsequent vicissitudes would have to be explained primarily
in terms of accidents and external factors and the first half of 1931 would,
in this case, give us a fair idea of what the rest of the depression would
have been without them.
2. Incidents, Accidents, and Policy in Germany. — During the first
half of 1931 one of those political events occurred which, at least in their
precise form and timing, we cannot but consider as purely fortuitous,
whatever our theory of history may be. Suddenly bursting upon an
unprepared world, the plan of an Austro-German customs union, natu-
rally interpreted as the forerunner of political union, was not, when pro-
posed in March, met by either of the two methods that would have,
temporarily at all events, prevented any further consequences, viz.,
resolute opposition or resolute acceptance, but allowed to develop into a
source of irritation and apprehension until, after protracted blundering
and floundering, it met its inglorious end on Sept. 3, which dealt a fatal
blow to the prestige of the last German cabinet that believed in action
within existing treaties. Under ordinary circumstances this affair
would, however, have remained in the political sphere and at first it
seemed that it would. General improvement continued, the Reichsbank's
gold stock increased somewhat, it was still possible to borrow abroad,
and even in the first days of May, 120 million marks of stock of the newly
founded Berlin Power and Light Company were placed with a foreign
syndicate. But then, the opposition to the customs union gathering
momentum, foreign balances began to make for home — no political lead
had to be given in order to effect that1 — and this was bound to produce
difficulties in the English and breakdown in the German market, which so
largely relied on them for current financing. In the eight weeks from the
end of May to the middle of July, Germany lost at least £J^ billion
1 An additional shock was imparted by the suspension of the Austrian Kreditanstalt,
the desperate situation of which became known to the public on May 11, 1931, and only a
few days earlier to the directors. But the importance of this event seems to the writer to
have been overemphasized. No doubt the failure of that famous institution, though
then only a shadow of its former self, was apt to make a profound impression throughout
the world, particularly because of its connection with the house of Rothschild, accentuated
as this connection was by the fact that the head of the Viennese branch of that family—
the private banking house S. M. v. Rothschild & Co. was not, however, directly involved —
happened to be its president. But the quantitative importance of the failure, however
serious for Austria, was hardly sufficient to exert any considerable influence on the course
of things in the world at large or even in Germany. Declared losses — swallowing up, it is
true, both capital and reserves — were about 20 million dollars.
THE WOKLD CRISIS AND AFTER 931
marks — not counting the credit of 100 million dollars extended to the
Reichsbank by the Bank for International Payments, the Bank of
England, the Bank of France and the Federal Reserve Bank of New York
on June 25 — and the Reichsbank's stock of gold and foreign exchange
exclusive of the proceeds of short-term emergency credits fell dur-
ing the year by about 2,200 million marks to 530 millions (middle of
December) . 1
Some aspects of this situation having been dealt with in the preceding
chapter, we need only notice the broad outlines of the consequent
catastrophe of Germany's financial structure, which spread paralysis
throughout her economic organism besides destroying a number of
fundamentally unsound concerns.2 The Darmstadter and Nationalbank
was closed on July 13 after a run which followed upon foreign with-
drawals and the failure of an important concern financed by it, and a
general run on banks ensued. Banks were closed by government decree
and, even after reopening, did not resume unrestricted payments to
depositors until Aug. 5, while the decree (July 18) "against the flight of
capital" ushered in the period of exchange control, which was tightened
on Aug. 1 and has since become ever more stringent to this day. Another
leading bank, the Dresdner, had to be reconstructed by the federal
government without actually failing.3 Savings banks were in a still worse
position as to liquidity, but resumed payment on Aug. 8. Events on the
stock exchanges, though they were closed from July 13 to Sept. 3, were
less catastrophic than might be expected. There were shares which did
not fall by more than 10 per cent within the year, and after the shock
imparted by the devaluation of the pound there was some recovery during
the rest of the year.
1The "standstill agreement" of Aug. 19, effective from Sept. 1, following upon a
period during which Germany discontinued her foreign payments — first without and then
with the informal consent of her chief creditors — of course, regularized but did not stop
the outflow of gold and foreign exchange, which amounted to over 500 millions in the last
8 months of the year. Payment of foreigners' mark balances and various lacunae in the
standstill agreement caused further loss, which it is difficult to estimate but which was
partly made up for by the favorable balance of trade.
2 Losses on ill-conceived ventures did not greatly affect creditors in the cases of the
Karstadt and the Linoleum concerns, but the case of the Blumenstein concern, which led
to default on the 1 million-pound loan of the Bank ftlr Textilindustrie, was more serious.
Still more so was the bankruptcy of the Norddeutsche WollkUmmerei, which owed 240
million marks, the Darmstaedter Bank being the chief sufferer. This bankruptcy occurred
in June, as did the difficulties of the Nordstern insurance company, which was, however,
saved. October brought the breakdown of the Schultheiss-Patzenhofer concern, which,
though a brewery, had entangled itself in enterprises in the fields of cement and textiles.
3 Also the Schroederbank suspended payment on July 20. Otherwise, there were no
major bank failures, though of course it is impossible to say what would have happened
without the "bank holidays" and the subsequent restrictions.
932 BUSINESS CYCLES
That business transactions were for a time not only hazardous and
unprofitable but in many cases impossible, stands to reason. All the more
important is it to stress the exceptional and extra-economic character of
the circumstances which caused the monetary mechanism to produce such
complete disorganization. The case teaches nothing about the working
of the gold standard and the merits or demerits of the classical or any
other theory about price levels, interest rates, and gold movements. All
that the political factor had put out of gear.1 Foreign ^credits had
financed German business to an extent that would, but for the political
factor (Chap. XIV, Sec. C), be inexplicable and is altogether unique.
They left Germany not under any of those stimuli which are inherent to
the mechanism of international finance but in utter disregard of them.
It follows that, since the efflux of gold and the liquidation enforced
thereby did not mean what they do mean in the ordinary course of com-
mercial fluctuations, there was no reason to react to them in the usual
way, viz., by restricting credit and raising the rate of interest, and that the
natural thing to do was to try to replace the working capital that was
being torn out of the German organism by an expansion of Reichsbank
credit and to leave the foreign value of the mark, for the time being, to
itself. Closure of the banks and many other embarrassments could also
have been avoided by this method, which, involving nothing but the
substitution of a fiduciary basis of credit for the gold basis that had been
destroyed by extra-economic events, would in no sense have been
"inflationary." Many observers believed at the time and most of them
believe now that the government and the Reichsbank were completely
blind to this and that, actuated by an uncritical belief in an outworn
fetish and oblivious of anything else, they clung to the gold standard and
precipitated and intensified the catastrophe by punitive measures. It is
suggested that this was not so.
True enough, government and Reichsbank were in no mind and in no
position to declare for, and to act consistently on, an "expansionist"
policy. They were not prepared to face the domestic and international
obloquy which, at a time when abandoning the gold standard was not yet
made respectable by the example of England and America, would in
that case assuredly have burst upon them. Fears that the foreign value
of the mark would be attacked by speculators and fall to lows which
would then tend to induce real inflation, and that at home, the fetish
once removed, there would be no halt to really inflationary expansion of
the circulating medium, were not unjustified. Respect for treaty
1 Given that fact, it may, of course, be argued that that mechanism cannot be relied
on and that a suitable policy and theory of precisely that situation must be worked out.
This is what — in a nutshell — the rest of this paragraph attempts to do. We are concerned
only with diagnosis. The writer has no intention of arguing for "classical" views.
THE WORLD CRISIS AND AFTER 933
obligations, promises, and recent enactments counted for something.
And, what perhaps weighed most of all with the leading statesman, that
remedy might have barred the way toward fundamental normalization,1
which had been embarked upon with such difficulty and was to preclude
all possibility of the recurrence of similar situations. He, therefore, kept
to the policy inaugurated before — the emergency decree of June 5 marks
another step — dealing with every emergency that confronted him in the
hope that it would be the last of its kind. Nor was this hope without
justification. In June it seemed to come true: President Hoover's
moratorium plan, though in its nature no remedy for the business situa-
tion of the moment,2 might have brightened the atmosphere and put a
stop to the influence of the political component of the situation. If it
had, Germany's position would have been all the better for having
adhered to financial decorum. It did not. But when that became clear,
the price exacted by official adherence to an orthodox policy had been
largely paid.
This, however, expresses but one aspect of the policy of those months.
Closer scrutiny shows that the other, the expansionist one, was by no
means absent. We start by recalling the general features of the inter-
national business situation down to the end of May. In accordance with
them, money rates fell everywhere — the Dutch bank rate was reduced to
2 per cent as late as May 16. The unrest among short balances, which
put an end to this, was originally confined to Germany (Austria and
Hungary being affected in sympathy). At first the Reichsbank did not
react at all. It was only after having lost about 600 millions of gold and
exchange from June 1 to June 11 and when it was faced with panic
demands on the twelfth and the thirteenth, that it raised its rate from
5 to 7 per cent. In doing so it acted on the hypothesis that this would
induce member banks to use their own reserves and impress at least some
of the components of that demand. Tension, in fact, relaxed and on the
strength of the hope raised by the Hoover moratorium no further restric-
tions were resorted to. The Reichsbank notes were, owing to the credit
of 100 million dollars mentioned above, still covered to 40.1 percent at
1 Nobody can say whether it would have. It may be plausibly argued, especially from
comfortable combination room armchairs, that a bold announcement of a policy of expan-
sion would have made normalization easier. Given German conditions, the writer does
not share any such belief. But it may well have been a fifty-fifty chance.
2 The plan, proposed on June 20, met with French opposition and was not definitively
adopted until August 11. To this delay it has become a tradition to ascribe its failure to
take effect. But to the writer the case seems rather to illustrate the limitations of the
"psychological" factor. The plan was certainly calculated to remove a cloud from the
horizon. But to the individual businessman, who the writer believes is less swayed by
irrational moods than is commonly believed, it held out little promise for the moment.
This is the reason why it produced nothing but an ephemeral stock exchange boom.
934 BUSINESS CYCLES
the end of the month, in spite of the new wave of withdrawals that was
incident to the rumors about the Darmstaedter Bank. With the crisis
gathering momentum and gold holdings falling precipitously, the Bank
indeed went to 10 (July 15) and 15 (Aug. 1) per cent — from which it then
receded on Aug. 12 (to 10 per cent), on Sept. 1 (to 8 per cent), and on
Dec. 9 (to 7 per cent) . But at the same time it did replace the gold asset
by bills and advances. Excluding the bills that were sent abroad for
rediscount, it held by Nov. 30 over 4.2 billion marks of bills, treasury
bills, and advances (Lombard), as against less than 2 at the end of May.1
This was made possible by the suspension of the clause concerning the
40 per cent cover of bank notes and by the foundation (July 26) of the
Acceptance and Guarantee Bank, which merely served to create financial
paper for rediscounting. Thus the authorities went as far as was com-
patible with upholding the foreign value of the mark — further, in fact,
for that value was upheld by measures that rendered it even then largely
nominal — and less pressure was exerted on business than is commonly
believed. That the severe slump in production and general business
which occurred in the third quarter of the year was primarily due to that
monetary crisis we do not by this statement call into question; but it
was due to the whole complex of causes and consequences of that crisis
and not to central bank action or interest rates in particular. On the
contrary, these played no role among the causes and but a minor one
among the consequences.
Neglecting minor difficulties, weakness of the mark in foreign markets
among them, we may say that the stage was set for some improvement
when England abandoned the gold standard, which for Germany was,
of course, a serious matter.2 The question why Germany did not follow
suit like the Northern countries has never been answered quite satis-
factorily.3 Perhaps a partial answer may be found in the consideration
that some of the effects of devaluation can also be secured by scaling
1 Liabilities of the six big Berlin banks decreased by about 3 billions during the same
period.
2 In some industries the flow of orders from England suddenly stopped and many
cancellations were received. Immediate effects were, nevertheless, not great. Ulterior
effects are difficult to estimate. We will pass by the stock exchange and banking events of
September, which also were among the consequences and eventually led to Germany's
application for the convocation of the Young Plan Committee. Owing to help extended,
especially from the Federal Reserve Bank of New York, the Reichsbank weathered the
storm comparatively well. At the end of October its notes were still covered to nearly
30 per cent.
8 The writer understands that it has been answered, by a man who ought to know, to
the effect that England "would not consent." But what could she have done? Nor
would it be satisfactory to repeat the string of arguments glanced at above. For England's
example made all the difference, and linking the mark to the paper pound would have
staved off the danger of disorganization through inflationary fears.
THE WORLD CRISIS AND AFTER 935
down all prices and incomes (see below, Sec. E). It might, hence, have
been argued that government would only have to go on with its policy
in order to achieve a similar end by another route. This is in any case
what they attempted to do by various measures, among which the
emergency decree of Dec. 8, reducing wages and interest charges,1 was
the most important. Whatever may be thought about this — the writer
is far from wishing to defend it — it was not simply nonsense nor, as can
be easily inferred from the similarity with devaluation, was it deflation
in any ordinary sense. And no dire consequences followed. Production
during the first quarter of 1932 was indeed at a much lower level than
during the first quarter of 1931 — although by less than the first quarter
of 1931 was below that of 1930, in spite of the greatly increased impedi-
ments of international trade, which reduced value of total exports to
little more than half of the 1929 maximum — but decline had come to a
stop, the Berlin Institute figures corrected for seasonal being for those
3 months (1928 = 100): 52.9, 55.1, 56.7 (and a little higher for the second
quarter). There was a relapse to a new low in July and August, after
which the upswing definitively set in.
These figures have been mentioned because they bear upon the remain-
ing question, whether that upswing can have been due to the different
policy that was adopted in the second half of the year. The negative
answer they clearly suggest — there is no earthly reason to believe that
the movement along the bottom in the first two quarters should, barring
accident, have been followed by anything else but recovery — is supported
by a survey of what was actually done. Some relief was given to the
taxpayer by the not uninteresting measures embodied in what was termed
the Papen plan,2 and more important than the actual relief afforded was
the spirit of this policy, which aimed at nursing the system instead of
harassing it incessantly. A public works program was also inaugurated,
subsidies — not altogether absent, however, under the preceding govern-
ment— were given for various purposes, especially for the repairing of
houses, and industrial construction was encouraged. But quantita-
1 The decree of Nov. 17 had already extended help to agricultural debtors, the situation
of whom was indeed becoming desperate. This decree may be characterized, as may some
measures later on taken elsewhere, as an attempt at regulated bankruptcy, again the same
thing which devaluation aims at by another technique. This measure and the anticipa-
tion of similar ones intensified the severe fall in the price of German bonds, whether payable
in marks or in a foreign currency, which occurred both at home and abroad in the last 2
months of the year. In New York German 6 and 6% per cent bonds fell below 20 per cent
by Dec. 4. But the incessant stream of restrictions also contributed toward making
balances in and claims on Germany all but valueless to the foreigner.
2 Its most interesting feature consisted in the issue of certificates for tax payments
which, as it were, embodied future tax remissions and were made eligible for discount at
member banks and the Reichsbank (Steuergutscheine) .
936 BUSINESS CYCLES
lively all that was palpably inadequate to produce a contracyclical move-
ment or even to lead out of depression. No observer of contemporaneous
developments has ever claimed that such homoeopathic treatment could
do that. Ample justice is done to that policy if the steadiness of the
subsequent recovery is, in part at least, recorded to its credit. To this
it seems in fact to be entitled precisely on account of its moderation,
which prevented hectic spurts with consequent relapses. But it would
be absurd to attribute the turn of the tide either to the mechajiical effect
of the few dozens of millions by which expenditure can at best have been
increased before the definitive upturn occurred, or the anticipations which
may have been induced.1 There was, it is true, the Lausanne agreement
(signed July 9, 1932), which reduced the German liability on reparation
account to 714 million dollars. But apart from the fact that we are apt
to exaggerate the effects on short-run business behavior of events that
do not immediately alter the data with which the individual firm works,
it is wrong psychology to argue that Germans must have felt relieved.
By that time most of them resented the idea of any further payments,
and this attitude had been strengthened by the international committee's
report (Dec. 24, 1931), which had stated Germany's complete inability
to pay reparations. The effects of the Lausanne agreement were, hence,
to say the least, doubtful and there can in any case be no question of their
having more than counterbalanced the really and directly depressing
effect that emanated from the steadily thickening fog of hostile tariffs and
quota. Comparative normalization of monetary conditions was, of
course, more important. The Reichsbank rate went down to 4 per cent
in the course of 1932. But this cannot be listed among external factors,
since it was the natural concomitant of the passing of deep depression
and, if anything, overdue.
3. Incidents, Accidents, and Policy in the United States. — In this
country the principles of a recovery policy began to take shape as soon
as the plateau on which business moved during the first half of 1931
gave way. This new dive in the third quarter of 1931, which was deeper
than we should have expected, was but to a minor extent due to the
unpleasant experiences this country had had in the role of a creditor
nation. Losses from defaults of foreign debtors on long-term loans and
from the depreciation or worse of other long-term investments, however
considerable, were small relatively to the size of the organism.2 More
1 The date of the decree which put the Papen plan into force is Sept. 5. If the Sep-
tember figure of the index of production was due to that modest though well-conceived
measure, Mr. von Papen has indeed wrought a miracle.
2 But such considerations as that the value of shares of foreign concerns was only 1.5
per cent of the total value of shares traded on the New York stock exchange and that the
former only participated to about the same extent in the total depreciation, which was
THE WORLD CRISIS AND AFTER 987
important embarrassments followed from the freezing or worse of short
credits to foreign banks, because this paralyzed the first line of defense
against the withdrawals of foreign balances. But until September 1931
there was hardly any sign that serious difficulties would arise on that score.
Repatriation of part of those balances was to be expected, because there
was little to do for them here. Throughout the summer this was, how-
ever, more than compensated by movements in the opposite direction and
by the middle of September the figure for total monetary gold stock
stood at a maximum of 5,015 million dollars: the golden armor of the
United States currency seemed to grow stronger as the Federal Reserve
System serenely sailed through those troubled months.
For the withdrawals which began on Sept. 20, the day on which the
Bank of England suspended gold payments, the writer has no other
explanation to offer but that this event suddenly convinced the world that
no currency was to be trusted and that it thereupon discovered weak-
nesses and inflationary possibilities in the American position, among other
things the possibility of a domestic run1 and the comparatively small
amount of gold that was "free" within the American definition. With
this the fact accords well that it was in the first instance the European
central banks which made haste to convert their holdings of American
exchange into gold. The gold exports to France, Belgium, the Nether-
lands, and Switzerland which ensued — the reserve banks lost 722 million
dollars (exports and earmarks) to Oct. 22, not in itself a very serious
matter — had not under these circumstances the classical but exactly the
opposite effect: there was a "crisis of the dollar" during which forward
dollars in Paris fell to a discount of from 5 to 10 per cent, President
Hoover's Second Plan (Oct. 6) becoming another argument, although per-
haps no reason, for feeling pessimistic about the dollar. But on Oct. 22
Mr. Laval arrived in New York. And the fact that his and Mr. Hoover's
professio fidei sufficed to stop the efflux and even to reverse it proves the
comparative innocuousness of the intermezzo. Moreover, French private
balances had by then been to a great extent repatriated, and the Banque
de France consented for the time being to discontinue withdrawals on
condition ( ?) that she would be spared a repetition of her English experi-
ence. Finally, the shock troops of international speculation garrisoned
roughly 58.5 billions as between September 1929 and December 1981, of course tend to
understate the real extent of the relative importance of total losses on foreign industrial
investments.
1 So far as gold is concerned, this possibility did not materialize to a significant extent
at any time during those 2 years. Total gold and gold certificates in circulation kept,
while displaying a bulge at the end of 1930 and the beginning of 1931, below the level of
1927, though apprehensions were entertained early in 1932. It was different with " money"
in circulation. In this item the big increase occurred in the second hah4 of 1931.
938 BUSINESS CYCLES
between Amsterdam and Zurich — 500 million dollars is perhaps an
exaggerated estimate, though it comes from good authority — had for
the time being a more interesting object to attack. In (February and)
May and June 1932 a similar wave of withdrawals occurred, though on
a smaller scale. Reversal in the second half of the year this time left
the country's total monetary gold stock slightly increased — release from
earmark and domestic production, etc., more than balancing exports —
while it had decreased, though only by 133.4 million dollars in 1931. J
This practically disposed of the rest of French claims.
With due respect to "psychology," no major feature of the conditions
in this country can possibly be ascribed to either those gold movements
themselves or such effect as they had (see below) on the structure of
interest rates. But another element asserted itself more and more as
1931 wore on, viz., the debt and especially the mortgage situation. Its
importance is not adequately measured by the number of bank suspen-
sions, although it was formidable: in 1931 2,298 banks had to pull down
their shutters or, more significantly, 1,702 from September 1931 to
January 1932, the period of the second epidemic,2 which was, of course, a
major factor in the general slump of the first half of 1932 and the then
soaring figure of business failures. Nor is it adequately measured by
the losses incurred by banks and other creditors on bad debts. Certain
classes of the latter, debts from installment sales for instance, behaved
remarkably well. But the strain and drain of the repayments that were
successfully made — Professor Fisher's debt-deflation — and the general
awareness of the fact that the value of collateral was impaired and the
net worth of so many people negative, partly enforced and partly sug-
gested restriction of operations all round, pressed on prices, decreased
employment. There is nothing astonishing in the fact that this situation
did not assert itself fully until about a year after the setting in of depres-
sion in our sense. It takes time (and an altogether abnormal burden of
debt incurred) for it to develop and for people to realize it and to cease
to act on the hope of speedy recovery. And it will then — this is not the
only case of its kind as an analysis of 1875—1876 would show — cause dents
and slumps in the interval in which deep depression should be giving way
to more gentle descent.
This being the diagnosis of the nature of American "incidents and
accidents," how far were their effects plus the fundamental processes of
depression influenced by recovery policy or anything else government and
federal reserve banks did ? That policy must, again, be defined in terms
of actual measures and not in ternus of disconnected, often — really as
1 Figures from Federal Reserve Bulletin or statements. The minimum in the total of
monetary gold occurs in June 1982, when it stood at 8919 million dollars.
2 From February to October 1982 death rate was comparatively low.
THE WORLD CRISIS AND AFTER 939
well as apparently — contradictory, always inadequate pronouncements.
What strikes us first is the handling of the international situation by
the coordinated efforts of the government and the reserve system with a
view, (a) to avoiding or mitigating breakdown of foreign credit structures
by direct help or by refraining from pressing American claims; (6) to relax-
ing tension by the moratorium of political payments; and (c) to minimizing
repercussions on the domestic money market. As regards the last point,
the aim was achieved to a greater extent than is generally believed. We
have seen that prompt action had, soon after the crash of 1929, reduced
money-market conditions to a state of ease which — for "natural" rea-
sons— prevailed through 1930 and still more so in the first part of 1931.
The rate of the Federal Reserve Bank of New York went down to 1.5 per
cent on May 7. Nevertheless, the reserve system bought from June to
August another 130 millions of governments, with the result not only that
member banks in the main centers accumulated considerable excess
reserves and that bond yields were forced down, but also that new financ-
ing revived.1 But the efflux of short balances after Sept. 20 put member
banks into debt again and also forced many of them to sell bonds. And
until the Glass-Steagall Bill (signed February 1932) removed the obstacle
to their borrowing from reserve banks, which lay in the scarcity of eligible
paper, by permitting loans on hitherto noneligible collateral, and the
obstacle to further open-market purchases of the reserve system by per-
mitting governments to be used as collateral for Federal reserve notes,
that is, through 5 months, there was indeed that "deflationary pressure"
of which so much has been made by some students of the depression. But
this pressure was altogether unequal to the inferences that have been
drawn from it.
Although the rediscount rate of the Federal Reserve Bank of New
York was, in obedience to the rules of financial tradition, raised to 2.5
per cent on Oct. 9 and to 3.5 on Oct. 15 and kept there to Feb. 26, when
it was reduced to 3 per cent,2 rates charged by member banks to cus-
tomers, which had reached a monthly minimum of 3.93 per cent for
September 1931, increased to a monthly maximum of but 4.72 for March
1932 (New York City; Federal Reserve Bulletin). This is the real test
of the severity of the pressure exerted, and on the strength of the most
common experience it disposes of the idea that there was discouragement
1 While this was going on, however, production and employment, from June on,
decreased strongly. Once more, such an experience, while not in itself sufficient to dispose
of certain theories about the efficacy of either rates or open-market operations, should not
be lightly brushed aside. In this case it cannot be urged that the power of the spiral was
unbreakable. For 5 months had preceded during which business had been "looking up."
2 Buying rate on acceptances was reduced on Jan. 12. Rediscount rate returned to
2.5 per cent on June 24, in sovereign disregard of the then gold outflow.
940 BUSINESS CYCLES
of industrial and commercial business. Immediately, however, after
the president's signature had been affixed to the Glass-Steagall Act, the
reserve system embarked upon the biggest open-market operation in its
history, buying $1.11 billion governments from March to August, driving
down rates all round and piling up excess reserves, more than balancing
the effect of the gold efflux in May and June which this policy produced.
This, of course, is the point at which a post hoc ergo propter hoc argument
is most likely to suggest itself. It need not, however, detain us, since
the redundance of the facilities created is obvious.1
The National Credit Corporation (articles filed Oct. 13, 1931), the
Home Loan Banks (bill containing the Glass rider2 signed July 22, 1932)
and the Reconstruction Finance Corporation (the original bill signed
Jan. 22, the extension, Emergency Relief and Construction bill, July 21)
represent the attempts made to mend what, in fact, were the most
important consequences of domestic "accidents and incidents," the
removal of which would allow the system to recover. The inadequacy
of the first measure to improve materially, let alone to save, the banking
situation — or else to make the economic system immune to it — is obvious,
as are the limitations of the second, which, however, within its limits did
something toward improving things in one sector of the mortgage embro-
glio. Both were far surpassed in importance by the third, which,
especially as extended by the Emergency Relief and Construction Act,
pegged a number of tottering structures, thus stopping up some sources
of infection from which cumulative disorders would otherwise have spread,
especially among banks and trust companies, railroads, building and
loan associations, insurance companies, and mortgage loan companies.
By Sept. 30, 1932, the grand total actually advanced — not merely
authorized — amounted to nearly 1.2 billion dollars, of which 185 millions
had then been already repaid, and the corporation had issued 750 million
dollars of 3.5 per cent notes, 600 of which were taken by the Treasury.
These few data3 suffice to indicate the aims and financial nature of the
measure during the first 8 months of the Corporation's life and to appraise
the kind and extent of the influence it can have exerted on the economic
processes around the lower turning point of the index of production.
Primarily intended as a support to banks and cognate institutions, and
as an agency to carry part of the burden of loans that were noneligible
1 This does not necessarily imply adverse criticism of that measure. To create redun-
dant facilities in order to provide for any requirements recovery might entail may be a
reasonable thing to do, even though not likely to induce an upward movement in a business
community that does not use the funds it already has or could have.
8 The Glass rider conferred an additional circulation privilege on government bonds
which extended the national bank-note issue, and would have been mentioned in the
preceding paragraph if it had been thought of sufficient importance.
8 They are from the Corporation's third report to Congress.
THE WORLD CRISIS AND AFTER 941
in the sense of the reserve bank legislation,1 its scope naturally included
the only type of big business that was seriously threatened, railroads.
The rationale of this is as obvious2 as are the considerable, if negative,
results: additional disasters were averted, but not much positive impulse
was imparted by it.
Also something was done under this scheme, especially in its extended
edition, for agricultural credit institutions — new regional agricultural
credit corporations were created, for instance — for financing the carrying
and marketing of farm products and so on, and this worked in with insti-
tutions and policies previously established for the benefit of the agrarian
interest. But as compared with the plight of a large part of the agrarian
sector, all that was done during those two years was surprisingly inade-
quate. Since 1929 the index of farm products at the farm had fallen by
over 60 per cent.3 Total figures of gross revenue from agricultural produci
tion — which, according to the estimates of the Department of Agriculture,
was about 9.4 billion dollars in 1930, somewhat less than 7 in 1931, and
about 5 in 1932 — and nation-wide indices of land values — as of Mar. 1,
that index was 115 in 1930, 106 in 1931, 89 in 1932 (1912 to 1914 = 100,
maximum of 170 occurring in 1920) — do not tell the whole tale. For a
minority, but a nonnegligible one, net income must have been negative,
and for a considerable minority net worth of the farm must have been
1 The Emergency Relief and Construction Act also, amending Sec. 13 of the Federal
Reserve Act, authorized Federal reserve banks to discount paper for individuals, partner-
ships, and corporations unable to secure "adequate" credit accomodation from member
banks (in our sense). This move to force member banks into lending freely was followed
up in the banking legislation of 1938, 1934, and 1935. In itself but an approach to Euro-
pean practice, it was part of a policy for which early American banking experiences must
have seemed ideal, and a concession to the theory — which is fundamentally wrong — that
banks hold a key position at the beginning of revival and that if their loans do not expand
this can only be due to their aversion to lending. But since the new powers were so very
soberly handled by the reserve banks, there is no need for us to go into this matter.
2 It was not so, however, for the man in the street, who did not see any relation between
his own fate and what he took to be just a contrivance by which a capitalistic-minded
government tried to protect fellow capitalists from the consequences of their own follies,
leaving the suffering masses to shift for themselves. This attitude, which asserted itself
very soon in Congress and elsewhere and was one of the first symptoms of the anticapitalist
storm that was brewing, was strangely neglected by the administration.
3 Nothing, of course, can be concluded from that fact alone or even from it plus the
fact that the index of the goods bought by farmers fell only by something over 30 per cent.
As we have seen before, this in itself may spell suffering but not breakdown and must be
considered in connection with the agrarian revolution of the twenties. But it spells
catastrophe when combined with the fact that in 1930 "39 per cent of the owner-operated
farms of the United States were encumbered with mortgages averaging 40 per cent of their
value. This means that about a fifth of the farms of the United States in that year had
mortgages representing over 40 per cent of their value." (Professor John D. Black.
The Agricultural Situation, January, 1933, Review of Economic Statistics, Feb. 15, 1933).
942 BUSINESS CYCLES
zero or less. Foreclosures increased rapidly and so did the proportion of
forced sales that were due to tax delinquency.1 Hence, it is clear that the
process of depression was in the agrarian sector allowed to go on even in
directions in which it would have been most easy and for a conservative
government, one would think, most imperative to stop it.
The Emergency Relief and Construction Act (Sec. 1, Title I) marks a
new departure in authorizing expenditure for relief and work relief, a
little more than two years after such a measure was indicated according
to our schema, or a little more than one year after the plateau of 1931
had crumbled. No measurable effects can, however, have emanated from
the 35.5 million dollars which were made available and the 14.2 millions
which were actually spent for that purpose to the end of September 1932.
The mentality which had such difficulty in reconciling itself to this —
anything but novel or radical — course of action and the persistence of
which is as curious as the violence of the reaction it produced, also asserted
itself in the incessant appeals of the chief executive for retrenchment of
public expenditure and increase of taxation (e.g., messages and pronounce-
ments of Dec. 1, 1931, of Jan. 8, of Mar. 8, of Apr. 4, and of May 5, 1932.) 2
In some cases there were special reasons for this, e.g., the weakness of the
dollar exchange in March and at the beginning of May 1932. But in all
cases account must be taken of the fact that that mentality was, until
it changed into its opposite, a datum of the situation which it was hardly
possible to modify to just the extent that would have seemed rational.
Under these circumstances a "budget crisis" was no matter of indiffer-
ence. And adherence in principle to what were considered sound methods
of finance was not unlikely to help recovery as well as to facilitate fiscal
normalization after the depression, provided it went not beyond what was
necessary to convince everybody that the budget would automatically be
balanced in future while for the time being expenditure was allowed to
unbalance it. This is precisely what the administration actually tried
to do. Owing to the insuperable prejudice that defeated the sales tax
(Mar. 24), 3 little came of the first part of the program, and the tax bill
signed June 6 and the omnibus economy bill signed June 30, 1932, cannot
have had any but reassuring "psychological" effects. But the second
part was all the more fully carried into effect. According to an estimate
used before, the net4 Federal income-generating expenditure amounted
1 Professor Black, op. cit., p. 10, puts that proportion as high as one-third.
2 It also asserted itself in the veto of the Garner- Wagner Relief Bill, which the writer
does not find easy to understand.
8 Without any undesired effects either during or after depression, 2 billion dollars could
have been raised by it with which to service and extinguish a depression expenditure of
from 10 to 20 billions.
4 Arthur D. Gayer, op. cit., p. 891.
THE WORLD CRISIS AND AFTER 948
to 1,748 million in 1931 and 1,646 million in 193& (calendar years).
There cannot be any doubt that this was the most directly effective part
of the government's policy — the real Emergency Relief — which only
gained in effectiveness by being coupled with official emphasis on those
" sound " principles that at first sight appear to be at variance with it.
The inference is that it prevented much potential disaster. Yet since
that expenditure — to which, of course, a very low multiplier would have
to be applied — did not stop the shrinkage in total outside debits, which
fell throughout the year, the inference seems reasonable that, although
by partly compensating the influence of Incidents and Accidents it
facilitated the turning of the tide, it did not turn it.
We will finally glance at the third epidemic among banks, which
belongs — as a belated installment — in the nexus of events we are now
surveying, although it ran its course entirely within incipient recovery,
lending it for a month or two all the colors of deep depression. It started
in November — the banking holiday in Nevada declared on the first of
that month may be taken as the starting point — gathered momentum in
January and February and was cut short by emergency legislation on
Mar. 9, 1933. The suspensions and holidays spread from Feb. 14 on
(Michigan; the first states to follow were Indiana, Maryland, Arkansas,
and Ohio) until on Mar. 5, almost complete stoppage of banking having
greeted the new president on his inauguration, they had to be ratified
by Congress. This time agricultural distress, more precisely the agri-
cultural mortgage situation, was not merely a contributory cause but the
main one, as is seen from the fact that the hurricane started in the agrarian
states of the middle and farther West and then moved to the East,
thus collecting the fine for the neglect of the agrarian plight. This
suffices to insert that panic into our picture. Its features are well known.
Distrust in banks, to some extent coupled with a distrust in the currency,
led to indiscriminate withdrawals of deposits and forced the banks in
turn to withdraw currency from reserve banks — member banks (in the
official sense) withdrew over 1.7 billions between Feb. 8 and Mar. 3 —
and from New York correspondents who lost almost 800 millions in
this way. On Mar. 2 and 3 alone, money "in circulation" increased
by nearly 700 millions, federal reserve credit outstanding by nearly
730. Loss of reserves and increase of notes outstanding reduced excess
gold reserves of the reserve banks by 1.1 billion dollars to 400 millions.
Domestic difficulties were increased by a simultaneous efflux of gold
from the country which — partly though not wholly induced by them —
amounted to over 270 millions in February and March. The New
York Reserve Bank had to rediscount with and to sell governments
to other reserve banks. After the banking holiday and under the pressure
of the Emergency Banking Act of Mar. 9 (amended by the Act of Mar. 24)
944 BUSINESS CYCLES
gold coins and certificates speedily flowed back, over 600 * millions return-
ing to the reserve banks before the end of March, so that their excess
gold reserves were, thereby and by the reduction of the amount required
to be held against notes outstanding, increased to 1,172 millions. In spite
of the restrictions to which gold movements had been subjected, the inter-
national position of the dollar was remarkably strong at the erid of March.
Member banks (in the official sense) holding about 90 per cent of all
member bank deposits were reopened by license on Mar. 15. By the
middle of 1933 the number of All Banks (including private banks under
state supervision and mutual and stock saving banks) operating under
license — no doubt many very weak ones among them — was 14,530.2
The immediate consequences of that panic, the new spiral it set in
motion, do not call for additional comment; but its ulterior consequences
cannot be too strongly emphasized. It completely demoralized all classes
and, by doing so, fundamentally changed the problem before the incoming
administration. Without it — and it was certainly an avoidable incident
— recovery policy would have been confronted with an entirely different
situation. As it was, the psychic framework of society, which till then
had borne up well, was at last giving way. Nobody for the time being
foresaw anything but continuing disaster, and everybody was resolved not
to put up with it any longer. The talk about impending revolution
presumably was nonsense; but it characterizes well the prevailing state
of the public mind, which, bewildered and exasperated to the utmost,
clamored for political action in redress of what every group in its own
way felt to be some grievous wrong. Politicians and "intellectuals,"
suddenly moved into a position of saviors and judges, had a rich keyboard
to play on. But the mentality of the country, the traditions of the
victorious party, the nature of the catastrophe that had to be dealt with
immediately, and the strength of the inflationist interests united the
majority of them on monetary expansion.
4. Deferring further discussion of the English case, we now return
to the question whether other American and German time series confirm
the location of the bottom of the fourth Juglar which above has been
determined mainly from the behavior of physical output.3 That they
otherwise behaved as we should expect, taking account of the incidents
and accidents just discussed, is obvious. Money rates in particular do
not seem to require additional comment — in Germany the prime bank
1 The total for all types of currency was $1,185 million, most of which went towards
canceling reserve credit — member banks' indebtedness decreased to $545 million by Mar.
29, when New York City banks again held excess reserves.
2 The total for the middle of 1928 was 25,941 and for the middle of 1932, 18,794.
8 It should be noticed that industrial consumption of electric power would do equally
well, but not total power production.
THE WORLD CRISIS AND AFTER 945
acceptance rate (Privatdiscont) was at 3.88 per cent in December 1933,
having steadily fallen from the peak in the second half of 1932 (7.95
for that half year), and in this country the third banking epidemic failed
to produce panic rates and only interrupted the downward course for a
short time, as the gold panic of September-October 1931 had done:
bankers' acceptance rate for 90 days' unindorsed bills was 1^ per cent
on Feb. 28 and only 2 per cent on Mar. 31, 1933, other rates moving
correspondingly.
The Harvard A-curve (speculation : index of prices of all listed stocks)
clearly indicates the trough for June- July 1932 and reacts well to the
incipient recovery throughout the third quarter of that year. Equally
corroborative is the behavior of German stock prices, the index of which
increased considerably in the second half of the year, climbing up to a
little over half of its value for 1925. But outside bank debits (and dollar
volume of department-store sales) continued their downward course,
with but an insignificant upward movement in the last (department-store
sales in the third) quarter, right into 1933, and thus at first sight seem to
cast doubt on our location of the trough.1 Considering, however, the
persistence of the fall in price level and the fact that bank debits were, of
course, particularly sensitive to the banking calamity, this does not mean
much. Moreover, incipient recovery is, as we know, compatible with
some further shrinkage of total dollar volume of business operations. It
would be compatible even with some further increase in failures —
although in this case maximum of failures actually occurred at the trough.
These and other symptoms2 may be likened to those symptoms of disease
which often show most markedly in convalescence. The real question
arises with respect to employment and prices.
In Germany the number of employed as per sickness-insurance
statistics was (monthly average) 17.6 millions in 1929, 16.3 in 1930,
14.25 in 1931, and a little below 12.2 in the first half of 1932. Then it
increased to October (12.9), to fall more than seasonally in December and
January — which, it will be recalled, is not contrary to expectation from
the experimental schema — before increase had gathered momentum.
The number of (statistically visible) unemployed rose by about 50 per
cent in the average of 1931, as compared with that of 1930, which was
1 Value of construction contracts (Dodge) increased, however, nonseasonally in the
third quarter, though they declined more than seasonally in the fourth. This is also true
of privately financed (mainly residential) construction. See J. B. Hubbard, The Con-
struction Industry in Depression, Harvard Business Review, January 1933.
2 In Germany, for example, household consumption even in physical terms definitely
declined in the third quarter of 1932, and the simultaneous increase in the output of
consumption-good industries at first went toward replenishing the stocks of wholesale and,
to a lesser extent, retail trade. As to monetary terms, turnover of cooperative stores per
member declined in the third quarter by 3.04 marks.
946 BUSINESS CYCLES
3.1 millions. The first quarter of 1932 displays the maximum of over
6 millions, the figure for the second quarter is 5.66 millions, and for the
third a little over 5.2. As noticed before and as we should expect, the
number of workmen employed throughout fell less and increased1 less
than production, and also lagged behind it. This is equally true of the
United States, where the behavior of employment was very similar.
The annual minimum, of course, occurs for 1932. More important
than this, employment in manufacturing industries begaij to increase,
slightly at first, in the second half of July. The Federal Reserve Board's
index then records some more than seasonal net increase for August,
though there was decrease in the automobile and allied, as well as in
the machinery, industries. Increase spread in September (the index
adjusted for seasonal then showed 60.3 per cent of the 1923 to 1925
average as compared with 58.8 per cent in August) and persisted to the
middle of November — when employment in the automobile industry
increased considerably — after which there was more than seasonal decline
in December and January. Unemployment (A. F. of L. estimate)
behaved accordingly. As in Germany, however, a new low point occurred
in the first quarter of 1933, which in this case is amply accounted for by
the banking crisis.
But price levels declined unequivocally through 1932 and for some
time after in all countries that remained on the gold standard and in
some that did not,2 and the failure of other series to display trough
and recovery in 1932 is primarily due to this fact. Since this might
prove disturbing to the reader who has followed so far but still holds on
to his habit of associating cyclical phases primarily and even causally
with movements of the price level, it is necessary to remind him that
the processes of recovery do not require that price level should first rise
or even cease to fall. It will be useful to recall how we should expect it
to behave at such a juncture, i.e., at the beginning of the recovery of a
Juglar (preceded by a Kitchin prosperity) which lies within a Kondratieff
depression. On the one hand, although there may be belated price
1 It will be recalled that this is due not only to statistical causes and to the fact that
workmen are not promptly discharged, so that at the beginning production increases
simply in function of decreasing short-time or underutilization, but also to the changes
in production functions (rationalizations).
2 Many individual prices, most of the world's staples among them, do, however, display
either minima and recovery or that flattening out (amidst "hesitations") of curves which
is the characteristic shape at the bottom of a cycle (as did also quantities). This may
be observed, for example, in the United States quotations (futures or spot price) of cotton,
rubber (which at its low in June was little more than 5 per cent of its annual average for
1918), zinc, lead, copper, iron, and scrap steel. Most of them lost part of their gain in
the last quarter — copper, in particular, because of the difficulties in the cartel. Even
the wool and petroleum curves flattened out.
THE WORLD CRISIS AND AFTER 947
reductions — in some cases induced precisely by producers' realizing that,
paralysis being over, reductions may now have some effect in stimulating
demand1 — the bulk of wholesale prices will, in fact, recover from panic
lows. But, on the other hand, such "correctional" movements are
superimposed on a fundamental tendency that works against their
effect on the index. We know that, and why, the price level should in
every neighborhood of equilibrium be at a lower figure than in the
preceding neighborhood, and in a Juglar within a Kondratieff depression
this tendency may result in its recovery phase ending up with a price
level below that of the lower turning point. Owing to the violence
of the break in prices during the preceding Juglar depression, this would
not have been likely to happen in this case even without the subse-
quent efforts to raise prices by political action. But that any rise
in prices that may have been due in reaction to depressive excesses was
slow in coming about and even that the fall in wholesale prices arid in cost
of living persisted for several months after the turning point of the cyclical
process is neither surprising nor a reason to question our dating, let
alone to date the cyclical trough February 1933. It should be observed
that this argument is independent of the fact that for this country
the bank holidays and the events that led up to them provide a special
and, according to our diagnosis, an "accidental" cause for that trough,
and therefore suffice to rule out that dating; for although this is true,
the fall in price-level graphs also went on in other countries.2
Particulars of price movements would merit discussion. We must
limit ourselves to a reference to the rich literature on the subject3 and to
the following remarks. First, barring the effects of monetary changes,
1 Or also by producers' so vigorously responding to an anticipated rise that instead of
it a fall ensues.
2 The influence of the epidemic among banks is not exactly measurable but very visi-
ble. The minimum of the cost-of-living index (National Industrial Conference Board)
and the minimum of the B.L.S. index of wholesale prices, both of which occur in February
1933, are obviously linked up with it, although these dents lie in a declining interval which
may be interpreted in the light of our schema: there was a rally in prices in the summer of
1932, which is absent or practically so in Germany — where, however, the index of sensitive
prices increased from 45.3 per cent of the 1913 figure to 53 per cent during the second half
of the year — but is well marked in this and some other countries, and which we may asso-
ciate with the Kitchin prosperity phase at the bottom of the Juglar. The B.L.S. index
shows a small rise in July, and many leading commodities advanced considerably in August,
while some that had advanced before declined. This up and down continued in September,
for which the index showed hardly any change. By the end of September decline had
become dominant, and for October the index was one point lower. This decline continued
during November and December, and the January figure, 4.2 points below that of August,
may already have been affected by the banking crisis. But the fact that there was much
greater price stability as compared with 1930 and 1931 is beyond doubt.
3 Mention should be made, in particular, of Professor F. C. Mills, Prices in Recession
and Recovery, 1936.
948 BUSINESS CYCLES
the fall in price level and in cost of living wa§ remarkably uniform. When
all qualifications on the score of comparability have been made, it is still
significant to note, for example, that the American and German indices of
wholesale prices yield, if 1929 is taken as base year, practically coincident
curves and that cost of living fell from 1929 to 1932 by about 22 per cent
in both countries. And this is not merely due to the influence of the
international prices which enter the indices.
Second, it follows from the argument of this book as a whole and more
particularly from what has been said in the preceding chapter that that
fall is not adequately characterized by being called an unforeseen disaster
or a catastrophe of the price structure wantonly wrought by monetary
factors or the vicious spiral — debt deflation, in particular — and the like.
No doubt these and other elements contributed to the violence and, in
many individual cases of raw materials and semifinished products, to the
extent of the drop from, say, the middle of 1931 on, when wholesale prices
had fallen by about 22 per cent of the 1929 average. But it has been
shown that a price level markedly below that of 1913 was what would
certainly have prevailed without the war and what, even with the war,
was bound to emerge in time as a result of the evolutionary mechanism
and as a consequence of and adaptation to the industrial revolution
of the age.1 The reader will think as he pleases about the desirability
or otherwise of letting this process have its way. But he should not
overlook — however he may appraise — its economic function and its
potential long-run results.2
Third, it has often been pointed out how differently different groups
of commodities were affected and how rigorously the price system was
changed thereby. The difference between the behavior of prices of raw
materials and the behavior of prices of manufactured products has
attracted particular attention and been held to have not only reflected
but also intensified the growing disequilibrium. All this is true to some
extent, but it does not tell the whole story either as to facts or as to infer-
ence. The minimum of the B.L.S. index of wholesale prices (February
1933) was 62 per cent of its value for July 1929. The National Bureau's
index of physical volume of total production3 gives for the minimum year
1 It should, however, be recalled that the fall in the wholesale price index from the
beginning of 1930 to the middle of 1932 was much sharper than the fall from the beginning
of 1873 to the middle of 1875, which was part of an almost steady decline that (not to count
the drop from the Civil War peak) lasted almost uninterruptedly from 1866 to the middle
of 1879.
8 No one who is not a socialist can, of course, wholly approve of these results. For
one of them would have consisted in large strata of independent or semi-independent agra-
rian, commercial, and industrial business being weeded out, hence in a long step toward
socialism. But this does not matter here.
8 See Professor Mills, National Bureau of Economic Research Bulletin, Feb. 20, 1933,
description of the index on p. 6. On prices see the Bulletin for Oct. 31, 1933.
THE WORLD CRISIS AND AFTER 949
almost exactly 62 per cent of the annual figure for 1929. But one com-
ponent, construction, contracted to 31.5 per cent, while the index of
building material prices was still 75.9 per cent, and the hourly wage rate
about 80 per cent of their 1929 values at the trough in February. Here
we have an obvious case of maladjustment. On the other hand, agricul-
tural raw products show no influence of the depression on output for 1931
and at most a small one for 1932: "the farmer accepts the cut" both
because he works under conditions of competition and because of the
technological peculiarities of his production. The opposite reasons do
not, however, wholly explain the fact that output of mineral raw materials
contracted a little more than total output; for there were, in addition,
elements of prime costs which failed to fall correspondingly, especially
wages.1 These elements — selling charges among them — acquire, of
course, increasing importance for manufacturing industry as we proceed
toward the finished article of consumption, and there is, hence, little to
be surprised at in the February figure of the price index for processed
nonfood consumers' goods (73.2 per cent of that of July 29). But many
manufacturing industries also "took the cut,'* for instance, petroleum
refining, food, tobacco, and leather product industries, paper and print-
ing, clothing, and house furnishing.2 Equipment industries did so to a
much lesser extent or not at all. But precisely in their case, reduction
of prices would hardly have stimulated demand.
These observations are in keeping with the view previously arrived at
on the subject of price rigidities. They are also relevant to the question
of the nature and consequences of the disruption by the depressive process
of the preexisting structure of (relative) prices. Our model does lead
us to expect dispersions — because of rigidities as well as for other reasons
— which spell disequilibrium. But it does not follow that every change
wrought by depression in the price system necessarily falls into that
category or that return to equilibrium necessarily requires the recon-
struction of the preceding system of relative prices. The contrary may
well be the case — an example is the price of copper, which the opening
of new sources of supply had turned into an untenable maladjustment that
would simply be conserved by uncritical attempts to restore either price
or income parities. Nor does it follow that every change in the price
system which does fall into that category necessarily impedes recovery.
It may facilitate it or be harmless. An example of the first possibility is
1 The farmer "accepts cuts" in two different roles: first, as a producer and, second, as a
laborer.
2 In a not unimportant number of cases actual price reductions were much greater than
would appear from indices based upon list prices. In a still more important number of
cases comparison between the fall in the prices of agricultural and the fall in the prices of
industrial products is rendered meaningless by the changes in the quality of the latter,
for which it would be necessary to correct.
950 BUSINESS CYCLES
afforded by any panicky and temporary drop in foreign-produced raw
materials; an example of the second, by the short-term rigidity of prices
of equipment goods, for recovery does not typically start from expansion
of real investment. Many particulars were, however, different in Ger-
many. We will merely mention that agrarian prices did not fall anything
like as much as in this country. In December 1932 they were about
65 per cent of the 1929 average, keeping about in step with the total
index and practically avoiding the catastrophic slide of the American
prices in 1931.
Adjusted demand deposits of reporting member banks in reserve cities
outside New York, after keeping up to almost the middle of 1931 fell
sharply to May 1932, after which there was a small increase that was
just about wiped out in the first quarter of 1933. "Country banks"
taken separately displayed but a decreased rate of decrease in the second
half of 1932. Net demand deposits of reporting member banks outside
New York City fell after the middle of 1931, first at an increasing then
at a decreasing rate, the curve flattening out and then slightly rising
in 1932. This is not exactly what we should expect but is accounted for
by the changes in the investment item, which increased strongly through
1930 and the first third of 1931, then fell to the beginning of 1932 and
increased again to almost the end of the year, while All Other Loans fell
strongly and almost continuously throughout and beyond. The index
of rate of turnover of demand deposits in principal cities, which we owe
to the Federal Reserve Bank of New York1 declined at a decreasing rate
near the end of 1932, when it began to rise again.
National Income produced, evaluated at 1929 prices, fell from 1930
to 1931 by a larger amount than it had fallen from 1929 to 1930, and from
1931 to 1932, the minimum, by a larger amount than from 1930 to 1931.2
Net corporate income (All Corporations except tax-exempt ones and life
insurance companies; before payment of income tax) became negative to
the amount of 2,850 millions in 1931, and for 1932 displays the maximum
loss of 5,200 millions. The number of corporations reporting loss was
greater than the number reporting positive revenue as early as 1930; in
1931 the relation was 284:176; in 1932 it was 366:80. Still more sig-
nificant are the figures of corporate accumulation, though the limited
value of such accounting items must again be borne in mind. Already
in 1930 it was negative to the amount of 4,110 millions. It was minus
6,040 millions in 1931, minus 6,550 millions in 1932, and minus 3,060
millions for 1933.3 "Business Savings" as measured by the Depart-
1 See its Monthly Review for June 1, 1935.
2 See S. Kuznets, National Income, 1919-1935, 193, p. 8, Table I, col. 4.
8 See S. Fabricant, Measures of Capital Consumption, National Bureau of Economic
Research Bulletin for June 30, 1936, and Statistics of Income.
THE WORLD CRISIS AND AFTER 951
ment of Commerce (National Income, 1920-1936, 1937) were minus
4,903 millions in 1930, minus 8,052 millions in 1931, minus 8,942 millions
in 1932, and only in 1935 reached a modest positive value for which year
Professor Kuznets* "net savings of enterprises" are still at minus 3,252
millions.1 The fact that the minima mostly (not for national income
measured in current dollars, however,) occur for 1932 is valueless for us,
because there can be little doubt that government action in 1933 is
responsible for that.2 Otherwise those data are, whatever their short-
comings, full of interest for us.
For the moment the situation that faced the incoming administration
certainly was, and for the future it looked, untenable. The unpopular
necessity of refinancing corporate business becomes particularly obvious
if, accepting again the National Bureau's method of correcting corporate
accumulation by subtracting the difference between depreciation on a
cost basis and depreciation on the basis of current prices — which still
leaves out of account inadequate provision for obsolescence — we realize
that the sum total of those accumulations for the period from 1919 to
1933 turns out to be minus 7,110 millions.3 This does not mean that the
sum total of cash items fell spectacularly. On the contrary, cash was
precisely the item that actually shrank least as compared with 1929 —
from about 7.5 to about 6.1 billions in 1932 — although omission of current
revaluation formally also kept up others; needless to repeat that this
phenomenon was wholly consequential and merely reflected but did not
cause the spiral. Nor does that mean that dividends fell as much as
earnings which (net income as percentage of capitalization) were minus
0.6 per cent in 1931 and minus 2.8 in 1932,4 only utilities, foods, beverages,
tobacco products, chemical and allied products, and (substantially)
printing and publishing, staying on the positive side throughout.6 Not
1 Professor Kuznets, op. cit., Appendix B, discusses these differences, the numerically
most important causes of which are the governmental dissavings and the adjustments for
changes in the valuation of inventories included in his estimate but not in that of the
Department of Commerce. In his series of net capital formation in business at 1929
prices (op. cit.> Table 13, p. 48, row II 1 6), which we have used already, negative values
begin in 1931 with 458 millions. 1932 displays the minimum of minus 2,600 millions
and the figure for 1935 is still negative.
2 That statement will, however, have to be qualified later on.
3 S. Fabricant, op. cit., p. 12. As has been pointed out in the preceding chapter, nothing
follows from that except the necessity for the individual firm to make accumulations which
may over time prove not to be accumulations at all. We could, for example, not argue
that it disproves any oversaving theory. On the contrary, the fact itself could equally be
invoked in verification of the theory that individual attempts to accumulate will precisely
result in the dissipation of still greater amounts (Ezekiel, Keynes).
4 Since net income includes other elements besides profits within our meaning of the
term, the latter were negative to an unknown but obviously very much larger percentage.
6 National averages hide many details which would shed light on our process. Thug
952 BUSINESS CYCLES
only stockholders in many corporations but, if considered as a class, all
stockholders were to a considerable extent allowed to live on their
capital. Thus already for 1930 negative accumulation ensued from the
payment of a total of net cash dividends amounting to nearly 5.7
billions, while all the net income that remained after tax payments was
less than 1.3. And in 1931 4.2 billions of dividends compare with a net
deficit plus taxes of over 3.2. While we shall think about the long-run
effect of this according to the theory of accumulation or saving that we
make our own, we cannot differ about the remedial or contraspiral effects
such a behavior must have had in the short run — however much they
may have been overcompensated by other factors — particularly since it
preceded the setting in of deep depression. Flotations of new secuities
accord with those contours. Corporate issues, foreign included, were
still 1,736 millions in 1931 but only 325 millions in 1932, while municipal
borrowing was active and the Federal government borrowed over 3
billions net.
Industrial pay rolls fell, of course, more strongly than employment
and arrived, in the middle of 1932, at about 40 per cent of the 1923 to
1925 averages. Per cent fall, less steep in 1931 than it had been in 1930,
then came to a halt. After declining substantially from the middle of
June to the middle of July, practically in all manufacturing industries and
many others (the main exception being the woolen-goods industry),
aggregate factory wage payments increased though subseasonally in
August and more significantly in September and October, after which
they dipped again to a new low point in February-March 1933. We
interpret as in the case of debits. The German development was
similar: the sum total of wages and salaries exclusive of pensions but
inclusive of salaries of public employees had, according to the estimate
of the Institut fiir Konjunkturforschung, fallen from its maximum of
about 44.5 billion marks in 1929 to about 41 billions in 1930, and con-
tinued to fall to 33.5 billions in 1931 and 25.9 billions, the minimum, in
1932. No fall occurred within the latter year, and the figure for 1933 is
slightly higher, but lower for the first half as compared with the first half
of 1932. In both countries, therefore, real wage bill (pay roll by cost-of-
living index) fell considerably. Average per capita weekly earnings in
the United States, as per monthly data of the Bureau of Labor Statistics,
declined from 1929 to 1932 by about one-third in manufacturing, only
insignificantly with public utilities and only by 12.5 per cent in retail and
the textile industry earned minus 6 per cent as early as 1980, minus 6.4 in 1931, minus 8 in
1933, which places it near the head of the list of losers. But if New England be excluded,
its place shifts to about the average. We see here with particular clearness the connection
of depression with the competing-down process.
THE WORLD CRISIS AND AFTER 953
wholesale trade. * The fall came about at a percentage rate that increased
to 193£ and continued in 1933 but at a decreasing rate, the total reduc-
tion in the end amounting to about 36 per cent in money and about 16
per cent in real terms.
Hourly rates fell, to the middle of 1933, but very much less — in some
industries, such as anthracite coal mining, not at all. In manufacturing
they declined from 59 cents in 1929 to 50 in 1932 and 49 in 1933.2 This
would yield a gain in real terms, and so would to a lesser extent the course
of money wages of unskilled labor as recorded by the Bureau of Public
Roads. But no estimate which aims at a single figure of nation-wide
significance can possibly yield a, fall in real rates. In Germany, available
quotations of hourly rates being official "tariff" figures, statistics may
somewhat understate the fall in money wages. But they were, at the
end of 1932, at about 78 per cent of the average of the maximum year
1930, which would make a decline almost exactly equal to that in the
cost-of-living index. Now two things are obvious from this behavior of
hourly rates. First, they cannot have been a factor in starting the
depression, whatever the theory we may entertain on the subject: their
fall cannot, because they kept up well at first and only reacted to a
depression already in full swing; their previous rise cannot, because as
we have seen before, it was altogether inadequate to produce that result.
Second, if there is anything at all in the view which has been discussed in
the preceding chapter, viz., that the long-run level of American money
wage rates, as distinguished from their cyclical variations, was "too
high" in the sense that it was partly responsible for the unemployment
of the twenties, then it is clear that the fall which occurred during the
depression was inadequate to correct that level, although the latter might
have been corrected by a subsequent rise in prices occurring without an
increase in wage rates.
But it is more difficult to say whether wage rates, by behaving as they
did, intensified or alleviated the depression. Since the dominating
factor in the short-run situations, especially of "deep" depression, is
the downward shift of individual firms' "demand curves" for labor, and
since many of them no doubt become less elastic in the process of shifting
downward, it is not only likely that actual reductions failed, for the time
being, to call forth additional demand for labor sufficient to raise the
total wage bill above what it otherwise would have been, and that greater
reductions would have still more completely failed to do so, but there
1 Even the figure for manufacturing is but an average with a very considerable disper-
sion, a fact which must be taken into account in judging effect. The figures for other
industrial groups vary widely from that and between each other — the one for bituminous
coal is 45 per cent.
2 Data of the National Conference Board.
954 BUSINESS CYCLES
must also have been cases1 in which reductions of rates simply resulted in
a decrease of total output and employment.
It will be seen, however, that this argument progressively loses force
as the system approaches the recovery point and that beyond it the oppo-
site conclusion suggests itself. Then our question admits of a much
more definite answer. We still have the same classes of cases before us.
But their relative importance changes when "demand curves" for labor
tend to shift upward and to become more elastic. Resumption or
expansion of operations begins, as we know, in individual spots. It
certainly did so begin in the case before us, so that the effect on prime
costs of individual firms is all that has to be taken into account. Firms
try to resume or to expand operations in a situation which, while no
longer discouraging, yet does not offer those enticements — profits in our
sense or any of those gains which are induced by the emergence of profits —
that later in the cycle may make moderate variations of wage rates a
matter of indifference. They are likely to calculate closely. Even in
the short run they have, particularly if starting afresh after a shutdown,
some latitude as to the combination of factors that they are going to
adopt. The prevailing cheapness of money will give them a slant toward
mechanization, which may be intensified by an increase and counter-
acted by the previous decrease in wage rates. Hence pay rolls are likely
to increase faster in the absence than in the presence of an increase in
wage rates, as long as there is abnormal unemployment. Thus it seems
permissible to infer, not only that such fall in rates as occurred facilitated
inception of recovery, but also that a stronger fall would, at least in the
American case, have facilitated it still more.2
D. The United Kingdom, 1931-1938. — In the preceding section
English developments were allowed, first partly and then completely, to
drop out of our picture because justice could not have been done within it
to certain features peculiar to them. We now turn to a discussion of the
few points that are relevant to our subject, and will at once carry our
survey as near the present time as possible. Throughout, it must be
kept in mind that England suffered from the repercussion of practically
everything that went wrong anywhere in the world and that she was a
chief sufferer from the general " incapsulation " incident to the world
crisis, though she was also the chief beneficiary of the fall in the prices of
raw materials. She reacted to all that by abandoning the policy of the
1 Various possibilities of this type have been pointed out by Mr. Harrod in his review of
Professor Pigou's Theory of Unemployment, Economic Journal, March 1934, p. 28.
2 There is no paradox in holding that a given event may intensify depression and, after
depression, facilitate recovery. Nor should there be any difficulty in disposing of the ques-
tion that is likely to arise if the above be read by itself, viz., how "demand curves" are to
start shifting upward if "incomes" be kept down.
THE WORLD CRISIS AND AFTER 955
Gold Standard Act, by tightening the economic bonds between herself and
the Empire (and some other countries) and, last but not least, by speeding
up that shift of her resources toward production for the home market, of
which the most important result, as well as symptom, was the Building
Boom.
1. The event that made England drop out of the line followed by
our narrative — or would, at all events, have caused the reader to think
that she did — was the suspension of gold payments by the Bank on Sept.
20, 1931, sanctioned the next day by the Gold Standard Suspension Act.
The world depression had merely made it more evident than it had been
before, that, under the social and economic conditions prevailing at home
and abroad, the maladjustment caused by the return to the prewar gold
parit would not disappear of itself and that, those conditions being
what they were, there was little if anything to be gained by fighting for
that parity, while at the same time the sacrifices and internal struggles
this would entail showed in their true dimensions. This being so, it is
ex post easy to understand that abandonment of the gold standard was
really a foregone conclusion when the state of the Bank's stock of gold
came in to play the role of Peel's potato disease.1 It had been well main-
tained, thanks to exceptionally favorable circumstances into which we
need not enter,2 during most of 1930 and in fact replenished after Septem-
ber 1929. But in November 1930 the efflux set in which was to prove
decisive. Government and Bank acted in a way which easily lends itself
to hostile comment both from advocates and from opponents of aban-
donment, but which, whatever the actual intentions may have been, was
"objectively" eminently wise. In order to take the plunge, and to
reverse the policy of the Gold Standard Act, with an unbroken front and
1 The analogy goes pretty far. The Irish potato plague and the distress caused by it
may have impressed the Duke of Wellington and other members of the Peel cabinet; but its
effects were not remedied by the repeal of the import duties on cereals but by direct relief —
the "Queen's pay" — which would have been no less possible had protection to agriculture
been retained. Similarly, the loss of gold did not absolutely force the hands of the govern-
ment or the Bank. But it provided an argument that facilitated the difficult tasks of tran-
sition. The writer has never been able to understand why some Englishmen should resent
this suggestion and insist on England's having been forced off gold by dire and ineluctable
necessities. Social and economic data, in England and the world at large, being what
they were, going off gold was an economically rational thing to do. There is no offense in
stating this, as far as the writer can see, and further than this he does not go. The author
who wrote a book about the "tragedy" of the pound must have a concept of that literary
genus entirely different from the present writer's. The ability of so shaping events that
every action seems to arise out of objective necessity and to embody the only possible
course to take, makes the unique greatness of English statesmanship.
2 Among them were gold imports from countries the currencies of which were in extremis
and which had to part with their gold, such as Japan and Argentina. Within the Empire,
Australia was in the same situation,
956 BUSINESS CYCLES
to disarm domestic criticism — of which, in fact, there was very little
and none that counted politically1 — it was first necessary to do for the
pound all that anybody could ask without being voted unreasonable by
public opinion. So the Bank borrowed 50 million pounds in this country
and France (Aug. 7, 1931) and the Treasury announced a similar trans-
action in the amount of 80 million pounds on Aug. 28.2 The Bank,
moreover, in July had already raised its rate in two steps from 2.5 to
4.5 per cent, and before that, in January 1931, had begun to reduce the
governments held by the banking department.
This raises the questions, very natural from the standpoint of defen-
ders of "orthodox" views on currency and banking, why both measures
were not taken earlier — those defenders would say "in time" — and why,
in particular, the Bank stayed at 2.5 per cent from May 14 to July 23,
i.e., during the really critical time. The rates of the New York Federal
Reserve Bank (1.5 per cent) and of the Bank of France (2 per cent) give
at best a partial answer. More important is the consideration that only
a very modest effect could have been expected in the existing situation.
But since some impression could no doubt have been made on gold move-
ments by vigorous handling of the tools of central bank policy, another
conclusion imposes itself: while perfectly ready to undergo sacrifices such
as are incident to borrowing, the Bank was not then ready, for the sake
of the pound, to exert significant pressure on the domestic organism.
The timing and dosing of the open-market operations mentioned, which
absorbed only funds that would have been idle in any case,3 is conclusive
on this point. Technically, however, the pound, unlike the dollar, was
"pushed off gold" and "went down fighting." And the sensation all
over the world — Continental bankers and economists could not have
been more completely stunned by the news that Providence had defaulted
on its bonds — was all the greater.
But in England there was neither panic nor — precisely owing to the
way in which the thing had been done or, if the reader prefer, had come
about — loss of "confidence," but rather a sigh of relief. However,
everything was done to keep up discipline and to prove urbi et orbi that
1 There were, no doubt, many individuals who, regardless of economic interests and
effects, sincerely grieved over what to them seemed a lowering of the flag. The word
ignominious actually found its way into the pages of the Economic Journal and in private
conversation it may have been said that "the Bank of England went bust." But these
feelings did not amount to anything. It might have been different if the Labor party had
been still in power. This was, however, avoided by admirable statesmanship on the one
side and no less admirable discipline and patriotism on the other: no party question arose.
2 Both credits were very promptly repaid — by the Bank on Oct. 81, 1981, and Feb. 1,
1932; by the Treasury on Mar. 4, Mar. 29 and Apr. 5, 1932.
3 This particular point, we have the satisfaction to state, is recognized by Mr. Hawtrey,
op. rit., p. 33.
THE WORLD CRISIS AND AFTER 957
this was not to be the beginning of loose finance and that there was no
warrant whatever for wicked Continentals to talk about South American
analogies. Bank rate was put up to 6 per cent on the day of suspension
and actually dwelt on that vertiginous peak for about 5 months — it was
first reduced to 5 per cent on Feb. 18, 1932, but 3 per cent was not reached
until Apr. 21; on June 30, 2 per cent came, to stay to the time of writ-
ing— dealings in foreign exchange were temporarily (till Mar. 2, 1932)
subjected to restriction,1 a temporary embargo on capital issues (relaxed
Aug. 30, 1932) was declared and, most important of all, the orthodox
principles of public finance were upheld by truly heroic efforts — an
economy bill among them, which was passed by the Commons eight days
after the suspension act — which availed to end the financial year 1931—
1932 with a surplus of 32.9 million pounds.2
From our standpoint it is necessary to notice that this policy, however
admirable some of us may think it was, exerted some pressure on the
economic process and, for the time being, tended to neutralize such
stimulating effects as the depreciation of the pound might have had.
This statement implies the admission that depressing effects may in
particular have emanated from the 6 per cent bank rate. Personally,
the writer doubts whether a single additional pair of boots would, during
those 5 months, have been produced had, other things being equal, bank
rate been lower. Still, since the rate might easily, i.e., without producing
per se any very catastrophic consequences, have been 3 per cent instead
of 6, and since it is no part of our theory of central banking to hold that
a difference as considerable as this is a matter of no importance to
producers' business, we will for argument's sake concede that its pressure
did extend beyond the open market, provided that the reader, in exchange,
be good enough, first, to observe on any of the familiar graphs that busi-
1 Those restrictions were, in April 1982, replaced by the establishment of the Exchange
Equalisation Account — misleadingly called a Fund — which serves the same purpose more
effectively and other purposes besides. Even from the standpoint of our subject, this
original gadget would deserve more attention than we can afford to bestow. It must suffice
to note that its operation may powerfully intensify or offset the open-market operations
of the Bank and be quite as effective in regulating the money market as these, though much
more difficult to observe. If the Fund simply bought and sold gold and foreign currencies,
the effect on member banks* cash and reserve with the Bank would be much the same as the
effect of gold movements; the only difference would, in this case, consist in that the Fund's
operations are planned. This effect, however, may be counteracted by concomitant sales
and purchases of treasury bills; but again, it may not. If the Fund sells, say, dollars and
deposits the proceeds with the Bank without relieving member banks of a corresponding
amount of treasury bills, a restrictive influence would be exerted. Hence, we have here a
new organ of central banking, which would be very important for us if we could go into
details.
2 The official result was a surplus of only 0.4 millions. But "expenditure" contains an
item of 32.5 millions for sinking-fund purposes.
958 BUSINESS CYCLES
ness was not more depressed during those 5 months than we should have
expected it to be in any case — the quarterly index of production of the
London and Cambridge Service first rose and then stayed up during
those critical months — and, second, to admit that if there was any
aftereffect later in 193£ it cannot have been very terrible, since by
August the building boom had set in. However, together with all
other measures taken — in particular, the balancing of the budget — this
policy no doubt substantially contributed to all those features of the
English process that make so striking a contrast with the course of
events in the United States: a speculative flare-up and the consequent
relapse were prevented, expansion was patiently awaited, secure foun-
dations were laid for the sustained and sober advance that was to last
with practically no interruption until the spring of 1938.
2. To complete this part of our sketch, as soon as it felt safe ground
under its feet, the Bank made a move toward monetary expansion, and
along with reducing its rate from 6 to 2 per cent, bought government
securities.1 The immediate object was preparation for the great refund-
ing and conversion operations of that year.2 But the ultimate effect was
to end the period of easy money at high rates — no reason to object to
this as a paradox — and to usher in the period of easy money at low rates.
Day-to-day money, for instance, which in the average of 1929 had been
at 4.57, fell to 0.68 for the third quarter of 1932, then was 0.66 per cent
in the average of 1933, 0.81 per cent in the average of 1934, 0.73 per cent in
the average of 1935, and from April of that year stayed put at 0.75 per
cent to the time of writing — the manometer of interest was paralyzed
almost as effectively as in the United States.
Those open-market operations, however, kept within sober bounds.
The annual average of government securities held in both departments of
the Bank, which in 1930 and 1931 had been about 295 million pounds,
was only 312 for 1932 and 335 in 1933. But this and the influx of gold
helped to swell the cash reserves of member banks (London clearing
1 In doing so, it not only offset, but during 1932 occasionally more than offset, losses of
gold. See on this and cognate subjects, S. E. Harris, British and American Exchange
Policies, I, The British Experience, Quarterly Journal of Economics for May 1934.
2 It must again be borne in mind that the money market and the banking policy were,
all along, dominated by the requirements and actions of the Treasury and by the mechanics
of public finance. Release of government loan dividends, provision for application money
for government issues, income tax payments, treasury preparations for heavy pay-
ments, and operations to facilitate all this or to offset the effects of it all were so important
that a history of the English money market during those years could almost be written in
terms of them. This is another important subject that must be dismissed in a note. That
open-market operation and the conversion of 2,086 million pounds of 5 per cent war loans
to 3.5 per cent (announced June 30, 1932) were, of course, necessary in order to remove an
obstacle on the road to cheap money.
THE WORLD CRISIS AND AFTER 959
banks, average in 1931, 182 millions; 1933, 212) and their investments
(London clearing banks, annual average of 1931, 301 millions; of 1932,
348; of 1933, 537; of 1934, 560; of 1935, 615; no increase in 1936;1 small
increase in 1937; decrease in the first quarter of 1938).
This increase in investments is, of course, reflected in deposits,
especially in their great stride in 1932—1933, which, however, did not
fail to be accompanied by a symptom of the truth that banks* invest-
ments tend to produce idle deposits; for the ratio of current accounts to
total deposits, which was about 54 in 1929 and then naturally fell — we
remember that there was no fall in total deposits in 1930; the fall in 1931
was insignificant, all the loss of gold notwithstanding, which shows again
that also in this case monetary stringency cannot have been a major
element in the depressive processes — reached its minimum (50) for 1932
and stayed near its minimum in 1933 (a little over 5 1).2 Total deposits
in the nine clearing banks were above the 1929 figure already in 1932 —
the further fall in current accounts that occurred in that year being more
than compensated by an increase in deposit (time) accounts — and, 1,914
millions in the average of 1933. The slight setback in business that
occurred in 1934 is reflected, in spite of continued increase (over the
year) of investments, in a small reduction, but the average for 1936 was
1,961 — 200 millions above 1929 — the average for 1936 over 2,100 and
that for 1937 still higher. No shrinkage but, on the contrary, further
increase occurred in 1938 (to September).
To the last quarter of 1935, advances contributed next to nothing to
this expansion of deposits. We have here the same phenomenon that
will have to be mentioned in the American and the German case — a
similar structural change in bank assets. Advances shrank sharply in
1932 and kept creeping along at about the same level for over two years
afterward. But they increased significantly — as they should within the
rhythm of our schema — from the fourth quarter of 1935 to the first
quarter of 1938, and for part of this period dominated variations in
deposits. The same is true of that part of discounts which consists of
commercial bills, if we may judge from the returns of the three banks that
publish these figures separately. Otherwise, the bulk of the portfolios
consisting of treasury bills, the discount item is not of much help to us.
3. Those effects of monetary management as practiced in England
which do not work via international relations may, therefore, be summed
up by stating — what will presently be verified by time-series evidence —
that it brought out rather than obliterated the features of the economic
1 This is the "trace" left of the "normal" movement of member bank investments in
the cycle, since 1936 may be considered as part of a Juglar prosperity.
2 In 1935 it climbed up to what was nearly the 1929 figure — another trace of the cyclical
regularity.
960 BUSINESS CYCLES
process which we should have expected to observe in the "normal" course
of things. Availing itself of the freedom gained by the disruption of
the galling gold ligamen, it no doubt protected that process from what
would otherwise have happened to it and avoided the painful operations
which, without going off gold, would have been necessary. But it
neither aimed at nor effected anything that could in any useful sense be
called an "inflationary" impulse, which in fact is not necessarily inherent
in either devaluation or depreciation. It (and concomitant policies)
not only did not aim at but, if anything, discouraged income generating
public expenditure and increase in rates of earnings or commodity prices.
Thus it was more remarkable and successful by what it refrained from
doing than by what it did. The effects which worked through inter-
national relations must be viewed in their place within a comprehensive
policy of reorientation.
a. The relief immediately given to foreign trade and conserved by the
stability of money wage rates and domestic prices is likely to be under-
estimated1 because it mainly consisted in warding off much of that
additional shrinkage which would otherwise have ensued, at least during
1932. There was, hence, no good reason for the disappointment expressed
at the apparent absence of any considerable immediate effects. Also,
the relief was in part counteracted by measures taken in other countries,
both in retaliation to and independently of the English move.2 As far as
these reactions did not consist in tightening quota and so on but in sym-
pathetic depreciation of currencies, it must, however, be emphasized
that they did not entirely defeat the English purpose even in the sphere
of international trade. For wherever the gold ligamen galled, its removal
would have helped English trade by improving the general situation in
the respective countries,3 and its removal coupled with that spending
policy from which England substantially refrained, would even have given
England a relative advantage in addition to the absolute one. Because
1 Methods are available for arriving at a rough estimate of that effect. For our purpose
it is, however, not necessary to enter into what would be a very laborious and expensive
investigation.
2 There really was a third category: without wishing to retaliate, foreign countries found
it easier to abandon the gold standard and to revise their promises to their creditors because
the procedure had been made respectable by the English example and no longer carried the
connotation it used to. The implications of this fact are most important for the general
theory of morals. Japan, for instance, who went off on Dec. 14, 1931, had no doubt very
good reasons for doing so. But, attaching much importance to her impeccable financial
record, she would presumably have hesitated to take the step (though it might have become
inevitable later on) without the precedent created three months earlier.
8 It is, hence, not correct to say, as was said at the time, that England tried, by abandon-
ing the gold standard, "to push the burden of depression on to other countries." Even
the element of truth, however, which is contained in this phrase, should in fairness read
that she tried to push back the burden that had first been pushed on to her.
THE WORLD CRISIS AND AFTER 961
of this and other reasons, it was perfectly logical to refuse to consider
stabilization of the pound at any particular value, and the permanent
refusal to do so would be an understandable policy.
Two other factors must, moreover, be taken into account. First,
as has been stated before, the great fall in the value of exports of manu-
factures had occurred from the last quarter of 1929 to the beginning of
1931. During the latter year, the rate of shrinkage rapidly decreased
throughout the three quarters preceding suspension; and in spite of
depreciation, the monthly average for 1932 was still lower, though no
doubt higher than it would have been without it. The average for 1933
(23.4 millions) was very little higher and even the average for 1936 (28.4)
was but 55 per cent of the average" for 1924, or 59 per cent of the average
for 1929. Only the first half of 1937 showed a vigorous upswing (season-
ally corrected figure for June: 37 millions) which then gave way to
relapse (June 1938: 30.7 millions, which was better, however, than the
values for April and May) . But this contour simply describes the course
of (English and foreign) cyclical phases and is what on a somewhat lower
level we should have expected to observe in any case. The conclusion
which follows for the degree of influence exerted by suspension is obvious.
Furthermore, part of the success attained must be attributed to adap-
tations to new conditions which would have asserted themselves, depre-
ciation or no depreciation: while the immediate effects of depreciation,
in 1931 and 1932, were particularly noticeable in the exports of textiles
to the Far East — British exporters making promptly concessions in the
foreign prices of their commodities1 — the increase which occurred in
1934, for example, was only to 15 per cent due to textiles, which tended
to drop into the background while other articles were coming to the fore.
This adaptation or shift was no doubt facilitated or conditioned by the
depreciation of the pound, but cannot be called simply its automatic
consequence.
Second, suspension was accompanied by England's definitive con-
version to neomercantilism. Her monetary policy is in fact, but the
complement of it.
6. The Abnormal Importations Act (Nov. 30, 1931; three lists of
emergency duties followed within the year) which marks the transition,
presented itself as, and in a sense undoubtedly was, an emergency
measure. It added to the effectiveness of depreciation in curtailing
imports and in relieving, on balance, the immediate situation. The two
measures combined account for the bulges in all wholesale price indices,
as well as in the indices of prices of foods, nonfoods, and materials, which
occur for the last quarter of 1931 and were eliminated or more than
eliminated by the middle of 1932. But excepting some effects on specu-
1 See Professor Harris op. cit., p. 487.
962 BUSINESS CYCLES
lative anticipations, depreciation did not immediately affect domestic
prices — or, for that matter, business activity — in any way other than by
the prices and quantities of English foreign trade.
The frank adoption of the principle of (moderate) protection by the
tariff act of 1932 acquires its historic importance and the glamour of vast
possibilities in connection with the 11 bilateral agreements between the
different parts of the Empire (passed by the House of Commons on
Nov. 3, 1932) which gave effect to the understandings arrived at in
Ottawa (July 21-Aug. 20, 1932) and crowned the efforts of more than
thirty years. It does not follow, however, that the importance of these
achievements is equally great for our subject and for the period under
survey. In Chap. VI we have seen that the immediate effects on the
cyclical process of the German Zollverein were by no means commen-
surate with what one might expect when judging from the long-run
developments for which it in part provided the frame. The same is,
mutatis mutandis, true of the case in hand and from our standpoint we
can dispose of it by three remarks.
First, although the new policy affected total imports of manufactures
considerably and still more many individual items, such as motorcars,
neither depreciation nor protection prevented values of total net imports
from recovering, in 1933 and 1934, fully as much as values of exports of
United Kingdom produce or manufactures. They increased more than
values of exports in 1935, 1936, and especially in the first half of 1937.
This means that the new policy did not prevent them from behaving
according to the cyclical schema.
Second, it is worth while to notice the contribution of depreciation
to the Ottawa success. It is hardly conceivable that things would have
been, comparatively speaking, so smooth if England had had to appear
in the role of the adamant creditor instead of appearing, as she was able
to do, in the role of the understanding friend who had just waived a
substantial part of his rights. Moreover, depreciation and monetary
management reestablished England's lending power, which the gold
maladjustment and the crisis had all but destroyed, and thus made her
again what she had been before, the giver of all financial delights.
This element we must bear in mind when judging the balance of
advantages and disadvantages that accrued to England from the depreci-
ation of her currency and from her methods of monetary management.
Both had much to do with the relatively good showing, during the world
crisis, of her income from foreign investments, which in turn mitigated the
impact of the depression on her domestic process. For instance, the
"income from British investments in Oversea Securities dealt in or known
to the London market"1 fell indeed from its 1929 figure of 212.4 millions
1 See the articles on the subject by Sir R. M. Kindersley in the Economic Journal. The
above figures are taken from the one in the number for December 1037, p. 654, Table IX.
THE WORLD CRISIS AND AFTER 963
to 138.3 millions in 1933 (minimum) and was only 164.4 even in 1936,
but that part of it which came from loans to Dominions, Colonial and
Foreign Governments and Municipalities kept up well during the crisis
years, though it continued to fall later for other reasons, and even the
total is much more comforting to look at than it would otherwise have
been.
Third, the Ottawa policy did not stop at the frontiers of the Empire.
The "sterling bloc" extended beyond and included several valuable
conquests. Tariff reductions and quota facilities for English commodities
were, for example, secured by treaties with the countries that had most
promptly linked their currency to the pound — Denmark, Norway and
Sweden — but much more than that resulted in the case of the convention
with the Argentine Republic (signed May 1, 1933, supplement Sept. 26).
In this case the remission of debt implied in the depreciation of the
pound, coupled with an additional loan, was made the cornerstone for
arrangements about the unfreezing of credits, commodity trade, and
control of foreign exchange dealings and so on, in a way highly advan-
tageous to Great Britain.1 It may take some enthusiasm to style that
arrangement as the granting of Dominion status to Argentina. But it is
an outstanding example of the able and conscientious handling of indi-
vidual sources of actual loss and potential gain which no doubt greatly
relieved the domestic cyclical process.
4. As stated above, English industrial processes displayed, besides
the general characteristics appropriate to the Kondratieff phase and the
special features characteristic of the downgrade of this Kondratieff,
also the effects of the reorientation of the English economy in response
to the change that had occurred in England's international position.
The first class of phenomena does not call for additional comment:
electric power (more and more completely developing into state enter-
prise), electric manufacture (wire, cables, installation, lamps, apparatus,
and machinery), motorcars, nonferrous metals, chemicals including
rayon, aeroplanes, and so on continued to progress without much inter-
ruption even in the worst year, the British-owned rubber plantations
being, within this class", the interest that suffered most during the crisis.
For some of the chief sufferers from reorientation, such as the coal and
iron industries, the situation was rendered still worse because the normal
process of innovation2 would have made them centers of depressive
symptoms without that. This was not so, however, with textiles and
shipbuilding, although the latter experienced a spectacular recovery in
1934 and again in 1936, lasting through 1937.
1 Cf. R. B. Stewart, Anglo- Argentine Trade Agreements, Canadian Journal of Economics
and Political Science for February 1936.
2 The chief instance of a major industry that suffered only, but suffered severely, from
the effects of economic "progress," is supplied by the railways.
964 BUSINESS CYCLES
In the second class of phenomena, the most important item, in fact
the item that may with but little qualification be said to have "carried"
recovery and prosperity was the building boom.1 It was a boom in a
consumers' goods industry, dwelling houses supplying from 1932 to 1934
over 70 per cent of the value of building plans approved, the share of
factories not reaching 8.6 per cent until 1936. And though mainly, it
was not only conditioned by that general shift toward production for
home demand and, as an aid to or symptom of this tendency, by the
impediments placed in the way of foreign investment or by the unpromis-
ing outlook for investment, either as to safety or as to rate of return, in
most foreign countries. Fiscal policy, the lead given in the twenties
by the policy of subsidies, the effect this subsidized housing had exerted
on the development of an efficient large-scale building industry, the stable
rates of earnings which in a time of falling cost of living left a large middle
stratum that suffered comparatively little from unemployment with a
margin to spend, the fall of building costs during "deep" depression, the
subsequent cheap money policy, industrial migration — all this power-
fully propelled either the demand for housing or the means of satisfying
it and thus helped to condition the boom in unassisted building. Some
of these favorable conditions were only temporary. Cost of living began
to rise (after relapse from the rise in 1931-1932) in the spring of 1935 and
then rose considerably to the summer of 1937. The index of building
costs (1924 = 100) which had been at 124.2 for 1930 and at 107.5 for
1931 was again at 124.5 in 1932 and rose to 165.4 for 1933. The mortgage
rate was slow in falling from its maximum in 1931. The first important
reduction occurred toward the end of 1932, i.e., after the boom had started,
and the effects of the further fall in 1933 together with the easing of other
conditions were more than offset by the increase in building costs. The
cheap money policy thus helped the building activity by shifting toward
it funds that would have had to be content, from the middle of 1932,
with less than 4 per cent elsewhere, and by thus increasing the power of
the specialized machine for the financing of home building rather than
by offering direct stimulation.
The activity in unassisted housebuilding as measured by the number
of houses completed without state assistance (Ministry of Health; the
figures are not quite complete; years are from Oct. 1 to Sept. 30) got
into its stride in 1930, the year after discontinuation of building under
the Chamberlain Act, 110,375 houses being completed as compared with
71,083 in 1929. The figure for 1931— the year of unprecedented and
1 See ante, Chap. XIV, Sec. D, 2. On the facts and the theory of the boom, cf. a
forthcoming book by Mr. W. Stolper; for the rise and policy of the Building Societies in
particular the article by Sir H. Bellman in Economic Journal for March 1933. Total
resources of building societies eventually rose to over 700 million pounds.
THE WORLD CRISIS AND AFTER 965
unbearable catastrophes1 — is 132,909 and that for 1932, the worst year
in our period, 132,886. Then follows an ascent at increasing rates to
1935 (283,453). But although declining, home building has kept going
strong to the time of writing. Even in the first quarter of 1938 the figure
of value of housebuilding plans approved was exactly the same as that
for the first quarter of 1937, the decline in total value of building plans
approved being exclusively chargeable to industrial construction, and the
figure for the second quarter of 1938 was but 7 per cent below that for the
second quarter of 1937. This boom is perfectly regular in the sense of
our schema. Its natural end will make a not less regular contribution
to what within that schema should be the depression phase of the current
Juglar.
Here we may notice the role in the cyclical process of actual and
prospective expenditure on armaments, which began to assert itself at
just about the time when the building boom had passed its peak: though
such expenditure is the almost unavoidable concomitant of neomercan-
tilist policy, it was the Abyssinian incident that caused it to step out of
the frame of the current budget and to become a new factor of the
economic situation. It means, on the one hand, the building up of a new
industrial structure for an armament industry of, at all events, new
proportions, which is the primary desideratum.2 Whether left to
private or effected by public enterprise, we have here a " conditioning "
task which may be perfectly adequate for carrying a Juglar prosperity,
even if we discount fantastic figures inspired by passing panics. On the
other hand, this means a break in the government's financial policy, of
which the abstention from anything like American spending — from
income generation by public spending — was one of the major elements
until them. What has, for reasons which the writer believes were justified
by results, been denied to public works is now perforce granted to
armaments. Effects may some day dominate the picture to the complete
extinction of cyclical contours, but so far this is not the case. For the
1 It is true that a boom in some small industry may start in deep depression, owing to
circumstances peculiar to that industry, without affecting the general picture, which may
continue to be one of deep depression. This is not possible in the case of a boom in dwelling-
house building, which presupposes conditions of demand, and entails effects on general
business, incompatible with deep depression. To speak of deep depression or of unrelieved
gloom amidst the beginnings of such a boom is an abuse of terms.
2 The English armament industry was always considerable, from the seventeenth
century on; but its size was a function of foreign rather than domestic demand. The
development during the World War was to a considerable extent undone during the subse-
quent 15 years. Exports of arms (including torpedoes and the like), ammunition, and
warships amounted in 1984 to only about 8.36 million pounds, and this figure characterizes
the size of the armament industry fairly well. There was, of course, a quickly expanding
aviation industry in addition.
966 BUSINESS CYCLES
fiscal year 1937-1938 the socalled special rearmament expenditure
financed by loans — which is additional to the Defense item of the
ordinary budget — was only 65 million pounds, and it would take much
more than that to induce what is euphemistically referred to as "expan-
sion," especially as long as vigorous taxation and economy — 21 millions
were in the same year economized on Civil Expenditures — are adhered to.
Some steadying effects on employment in recessions and depressions and
some pressure on the standard of life are all that can be immediately
expected.1
Budgetary deficits, if any, continued to be insignificant. So was the
increase in the National Debt, if we exclude that part of it which was
incurred in order to finance the gold purchases of the Exchange Equalisa-
tion Account. But it would be forcing open doors if we tried to prove
that no government "inflation" lent its aid to the steady progress of
those years.
5. Evidence from time series confirms the impression we have gathered
so far: neither new policies nor the numerous disturbances to which
England's economy was exposed, effaced the cyclical contour lines which,
taking account of specifically English conditions, we should expect from
our schema.
Whether we take the Board of Trade's index of production or that of
the London and Cambridge Economic Service, we cannot fail to be struck
by the gentleness of the descent into, and the steadiness of the recovery
from, the depression, which characteristically distinguishes the English
curve from any other, the Nordic countries which followed the same
monetary policy not excluded.2 This steadiness is as much in evidence
before as it is after the abandonment of the gold standard.3 Between
the middle of 1931 and the middle of 1932 there is a little hump which is
not absent but much less marked in some countries that remained on the
gold standard, so that it may, partly at least, be attributed to deprecia-
tion. The minimum, which occurs in 1932, is lower, however — the Lon-
don and Cambridge Service quarterly index has 77.8 (average of 1924 =
100) for the third quarter — and the recovery of the fourth quarter then
strongly continued in 1933, without any appreciable rise in prices, leading
1 The above paragraph was written in the late summer of 1988.
2 See e.g., Chart I on p. 256 of D. Westcott's article, previously quoted, in the Review of
Economic Statistics for December 1934.
8 That gentleness of sweep — in some cases approaching, until 1983, a horizontal — is
also observable in the majority of components, especially in Foods and Drinks, Leather and
Boots, and from the middle of 1980 on, in chemicals, textiles, non-ferrous metals. Engi-
neering including Shipbuilding and Mining displays a stronger downward slope to middle
of 1932, and Steel is the most cyclical series of all. The steady upswing in the latter from
middle of 1932 clearly falls into two segments, one covering 1933, 1984 and hah8 of 1935, the
rest covering the second half of 1935, 1936 and three quarters of 1937.
THE WORLD CRISIS AND AFTER 967
up to 102.5 for the first quarter of 1934. The general improvement of
things which is observable in that quarter was associated with a tem-
porary reduction of the income tax. In any case, however, and whatever
the shortcomings of indices may be, we have before us a broad bottom
extending over 1931 and 1932 which in time and shape exactly corresponds
to the idea we have about how the depression of a Juglar located as this
was should look, followed by a perfectly normal recovery phase which it
would be highly unreasonable to attribute to monetary policy as "the"
cause though, as we expressed it above, this monetary policy undoubtedly
"protected" it from the consequences that adherence to the gold standard
would have had. It is also perfectly en regie that recovery tapered off in
the course of 1934, and so is the vigorous upswing in the last quarter of
1935, which, had we a desire to press the point, could be identified as the
rise of the fifth Juglar.
The statement that recovery was perfectly normal would, however,
require qualification if we accept the testimony of the London and Cam-
bridge Service's quarterly index; for its figures for 1934, though higher
than those of 1924, are lower than those of any of the years from 1927 to
1929 and thus fall below expectation. But we know that this is due to
the inadequate coverage of that index. The annual index of the Service
gives 120 for 1934 and 126.3 for 1935, which is substantially above the
previous maximum of the postwar period (1929: 115.8). The completion
of the preliminary reports of the 1935 census of production supplies us
with confirmatory evidence. Data that may be said to present the
balance sheet of the great depression have been analyzed by the Service
and warrant the conclusion "that physical output in 1935 was about 20
per cent greater than in 1930. l Output per operative employed in All
Industries increased still more,2 engineering heading the list — then follow
nonferrous metals, textiles, paper, and building — which suggests the
presence of quite as strong a technological component as we should
expect to find at that cyclical juncture.
Related to this technological component is the fact that part of the
innovations the consequences of which worked themselves out during
the period, i.e., those connected with the general reorientation of the
English economy, involved change of location and, hence, migration.
This alone would account for an amount of unemployment supernormal
even for that phase of the Kondratieff .3 Moreover, the general factors
1 Special Memo. 47, A. L. Bowley, J. L. Schwartz, and E. G. Rhodes, Output, Employ-
ment and Wages in the United Kingdom, 1924, 1930, 1935, p. 27.
2 Ibid., p. 35.
8 The slightest unfavorable variation in employment is by a sector of the daily and
weekly press hailed as proof positive of the failing of capitalism and used for derogatory
comparison with forms of organization of the bolshevist and fascist type — fascism being,
968 BUSINESS CYCLES
which made for supernormal unemployment in the postwar period and
which we will not list again, must be borne in mind. On the high level
set by this frame, the fluctuations of unemployment in England were
" normal" in our sense. Taking the number of unemployed insured
males as an indicator, and recalling that in 1930 figures rose above the
one-million level of the twenties, we find them moving at or above 2 mil-
lions (practically) throughout 1931 and at nearly 2.4 millions during 1932,
which also in this respect was the worst year though prices were stable
enough. A sharp decline followed in 1933, which again tapered off
through 1934 and part of 1935 and then went on at an increased rate,
approaching the one-million line in the third quarter of 1937. In Decem-
ber 1937 a hyperseasonal increase occurred and a little above the figure
then reached (1.31 millions) unemployment remained through July 1938.
There is, barring external factors, no reason for expecting very great
improvement in the situation within the next two or three years. But
the armament or other spending programs may do something. Also
migration has got under way and may reduce the "level" of the shorter
fluctuations.
This behavior of unemployment is strictly within the cyclical schema
and reveals little, if any, influence from monetary policies. The descent
was not slower or smaller, the recent ascent not quicker or greater than
we should expect under English conditions, taking account of the burdens
imposed on the economic organism. But the inequalities as between
areas and industries cannot be sufficiently stressed.1 In June 1936, for
example, unemployment percentage in London, the South-east, the
in this point, commended even by writers not otherwise in sympathy with it. There is
point in this comparison, but it is worth while stating in what it consists. As far as changes
in the productive organism involve geographical shifts of large numbers of workmen
families, no administrative device could entirely avoid the emergence of what would still
be unemployment in the economic sense. Though "temporary," it might have to last for
years, although a vigorous administration might prevent it from showing by commanding
the workmen to do something meanwhile. It is true, however, that such periods need not
last as long as they are likely to last in a capitalist organization. If workmen can be ordered
about regardless of their will and immediately incur severe penalties in the case of dis-
obedience, the importance of this, as well as of other sources of unemployment, will natu-
rally be much diminished. No theories about vanishing investment opportunities are
necessary to account for that large increment of unemployment which owes its existence
to the fact that in modern capitalism the workman is a free and very powerful citizen.
Unwillingness to admit this palpable truth is one of the reasons why so many of us take
refuge behind theories which without that inhibition they would hardly think worth
discussing.
1 Cf. the highly instructive analysis by D. G. Champernowne, The Uneven Distribution
of Unemployment in the United Kingdom Review of Economic Studies for February, 1988.
It must be remembered that these percentages refer to all insured persons from which
English, unlike German, legislation excludes domestic servants.
THE WOULD CRISIS AND AFTER 969
South-west, and the Midlands was 7.3 per cent, while in the North-east,
North-west, Scotland, Wales, and North Ireland it was 18.7. In August
1936 the unemployment percentage in all industries was 12. 1;1 but in
tramway and omnibus service for example, it was only 2.9 and in ship-
building and port transport 30.5.2 In electrical engineering it was 2.6
per cent in the first of the two areas mentioned, and 6.3 in the second.
These examples suffice to show the range of differences, which conclu-
sively proves that much of the total volume of unemployment was
intimately linked to conditions peculiar to individual industries and that
a great part of the phenomenon must be missed by aggregative theories,
diagnoses, and remedies. This does not mean, of course, that "causes"
are not amenable to general formulation. On the contrary, these differ-
ences are the very results of our process of reorganization and elimina-
tion, which we designate by the term competing -down, including therein
the geographical rearrangement. We are faced both with states of
disequilibrium and with a movement toward a new equilibrium. Hence,
however serious the temporary difficulties and hardships, the problem is,
as far as that goes and barring the probably permanent social pressure
exerted on the capitalist machine, not in itself a permanent one. It has
nothing to do with structural peculiarities of capitalism — other than the
one mentioned in a previous footnote — and, in particular, nothing with
any inherent inability of the capitalist mechanism to attain equilibrium
or with any inherent tendency of it to establish subnormal equilibria.
The wholesale price index of the Board of Trade rose, after its relapse
in 1932, considerably in 1933 and then remained, first because of opposite
movements in prices of foods and nonfoods, then because of all-round
stability, about at the level attained until the third quarter of 1935, when,
agreeable to expectation, a general rise set in. This rise, in which prices
of materials were the dominant component, lasted to July 1937 —
exactly as long as it should have lasted according to the cyclical schema —
and for that month carried the index to 80.2 per cent of the 1924 average,
i.e., near the 1929 average (82.2). Cost of living (index of Ministry of
Labour corrected for seasonals) stayed longer on the level reached by the
relapse of 1932 and did not rise materially until the spring of 1935.
Then it attained a maximum of 92 per cent of the 1924 average in April
1938 and thus also came near to the 1929 average, which was 94. Though
regular as to fluctuations, this behavior both of wholesale prices and of
1 It fell to about 10 per cent in September 1987, at the high- water mark of prosperity
and with acute shortage of especially skilled labor in some lines and places. And this
10 per cent covers 5.8 for London and the South-east area and 16 per cent for the Northern
area, Wales and North Ireland displaying respectively 19.6 and 21.
2 C/. Sir W. Beveridge, An Analysis of Unemployment, Economica for November 1986
and February and May 1937.
970 BUSINESS CYCLES
cost of living is not quite what we should have expected from our model.
Recovery and prosperity should have done little more than put a tempo-
rary stop to the fall, and price-raising policies all over the world must
be held responsible for the rest. So far as the rising level was due to the
increase in the prices of imported raw materials, it obviously cannot have
contributed much to English recovery, although it may have stimulated
exports indirectly. The terms of trade ceased in 1936 to be as favorable
as they had been during the six preceding years.
Pressing the new Index of Average Weekly Wages compiled by the
London and Cambridge Economic Service, into a role which, as we
know, it is not quite qualified to play, we will note that money wage
rates, which had been very stable at the 1924 average until the middle
of 1928, J/£ of 1 per cent below it to January 1930 and a little over 98 per-
cent of it until January 1931, went on slowly falling through 1932,
reaching 94.5 for the last quarter of that year. Then weekly wages began
to rise, continuing to do so for the rest of recovery and through prosperity.
The maximum of 103.25 was not, however, reached until April 1938,
i.e., after recession had set in. They have remained at that level until
the time of writing. This, as compared with 1924, makes a gain in real
rates of about 15 per cent, or about 6 per cent as compared with 1929.
Thus the new policies reduced, but did not eliminate, the increase in real
rates which at that juncture our process would normally produce.
Total money wage bill (including salaries)1 was at a minimum for
1932 (2,223 million pounds, as compared with 2,251 million pounds in
1931 and 2,430 million pounds in the maximum year of the twenties,
1929) and not much higher in 1933 (2,269 million pounds) but both years
were above 1929 as to real wage bill — the total real income of wage and
salary receivers thus behaving in a way which again negatives the idea
of unrelieved gloom. The figures for 1934 and 1935 (2,364 and 2,457
millions) answer well to our idea of the behavior of money wage bills in a
normal recovery. The astounding steadiness of the English develop-
ments and the complete absence of "inflation" — so impressively asso-
ciated with the mildness of the downturn, 1937-1938 — is still more
obviously reflected in the behavior of total monetary home-produced
income minus government expenditure (the maximum in the twenties,
3,592 millions, occurred in 1925 as the cyclical schema would have led us
to expect), which was at its cyclical minimum of 3,138 (1929: 3,553)
millions in 1932 and recovered to 3,745 millions in 1935. London country
and provincial clearings tell the same tale, which is confirmed by new
corporate capital issues for domestic purposes — which did not really
revive until 1935 — and by the course of prices of industrial shares, which
1 These as well as national income figures are taken from C. Clark, op. cit., Table 39,
on p. 94.
THE WORLD CRISIS AND AFTER 971
rose practically steadily from the beginning of the third quarter of 1932
to the end of the third quarter of 1937.
E. The State-directed Economy of Germany raises questions of
economic and sociological principle which could be treated only within a
research program much wider and more detailed than ours. The little
sketch which it is, nevertheless, necessary to present will be carried to
the spring of 1938.
The outstanding feature is the rapid progress, practically without
relapse, toward full employment of resources in general and labor in
particular, in fact toward more than that: unmistakeable symptoms of
overemployment in our sense show in some industries before, in many
industries in, 1937 — among them measures to relieve shortage of labor.1
The Berlin Institute's index of total industrial production (1928 = 100;
1929 = 101.4; 1932 = 54) rose in the monthly average to 95.3 for 1935,
107.8 for 1936, 118.8 for 1937, 120.7 for the first quarter of 1938, 125.9
for April 1938 (April 1937 = 118.5), the index of what the Institute
classes as industrial producers' goods, from little over 40 at the beginning
of 1932 to 135 in April 1938 (equipment goods alone, from about 30 to
138), that of industrial consumers' goods, from less than 70 to 110. The
production of raw steel rose to almost 20 million tons in 1937 (United
1 The significance of what in our terminology is overemployment of labor (employment
greater than neighborhood employment) has been questioned on two grounds. First, it
has been pointed out that German statistics include in the number of employed, persons who
in the United States would be counted as unemployed. Absolute figures of employment
and still more of unemployment are, in fact, internationally incomparable. Although
comparison of changes is less misleading, we recognize this by stressing the obvious symp-
toms of the presence of a problem of shortage. If full employment were merely statistical,
there would obviously be no endeavors to cope with shortage (considering the level of wages
which precludes the possibility that it is cheap labor that is lacking) by allowing temporary
immigration, increasing hours, increasing the employment of women, reinserting older
workmen, and the like, and no measures would be taken about it, such as the decree of
June 1938, introducing general compulsory service for "tasks of particular public impor-
tance." The statistical decline of unemployment from 5.7 millions in April 1932, to 0.4 in
April 1938, hence, cannot be dismissed on this ground. At the end of May 1938, 195,000
industrial standard jobs were vacant.
Second, it is of course true that the rebuilding of the army, the party troops and party
activities in general, working and other camps and similar organizations for the young,
and workmen's holidays absorb a great many people. But some who are thus absorbed do
work which would have to be done in any case, such as traffic regulation and even industrial
work. Moreover, the increase in the employable population must be set against that
absorption, as well as the increase in women's and foreigners' employment and the higher
retiring age. Finally, the increase in the total number employed, from April 1933 to April
1938, was (insured persons including servants and salaried employees) about 7 millions,
that is, greater by about 2 millions than the decrease in the number of unemployed. The
amount of invisible unemployment in 1933, though considerable, cannot have been nearly
so much as that.
972 BUSINESS CYCLES
States: 51.7) and surpassed this record in the first quarter of 1938, reach-
ing an all-time peak in spite of the shrinkage in exports. Domestic
orders received by the machine industry for the first months of 1938 were
almost exactly seven times the monthly average for 1932, motorcars
produced (passenger and trucks) over six times. Electric power produc-
tion nearly doubled between 1933 and 1937. Value (current marks) of
industrial and commercial construction1 (1932: 0.6 billions) did not rise
at all in 1933, moderately in 1934 and 1935, and significantly only in 1936
and 1937 (to 1.8 in the latter year), and the value of dwelling-house
construction (1932: 0.8) percentually still less (to 2 billions in 1937).
If, nevertheless, the building industry worked at full capacity from 1937
and efforts had to be made to economize labor and materials, this was
due to public building (1932: 0.9 billions) which increased by leaps and
bounds to 6.2 billions in 1937. Total physical volume was then con-
siderably above — that of dwelling-house building about at — the level of
1928 and 1929, though it is difficult to say exactly by how much, because
of the change in the combination of cost factors that has occurred. The
Institute's index of building costs rose from 1933 on but in 1937 still
stood at only 76.8 per cent of 1928.
This is exactly the kind of performance that our model would have
led us to expect from unfettered capitalism. Very obviously, however,
capitalism was not unfettered. In trying to diagnose the nature and
effects of government leadership and control, we may discard the policies
which revolutionized the structure and organization of the agrarian sector:
vastly important though they are from other standpoints — in some
respects, perhaps of seminal importance for the solution of the farmer's
problem — all that need be noticed for our purpose is that the index of
agricultural production was in the average of 1934-1935 to 1936-1937
higher by 19 per cent than in the average of 1927-1928 and 1928-1929,
and that the index of agrarian prices in 1936 and again in 1937 was about
32 per cent above its minimum at the beginning of 1933.
We may also dismiss the policy of forcing exports by means of arrange-
ments about blocked marks, a direct subsidy, and bilateral agreements.
This policy, which helped to increase values (current marks) by over
40 per cent from 1934 (the annual minimum) to 1937, was dictated
by the necessity of providing exchange for the improvement of the
foreign debt situation and the purchase of raw materials. But it counter-
1The data about values of construction are taken from the semiannual surveys
(Deutschlands wirtschaftliche Entwicklung) published by the Reichs-Kredit-Gesellschaft.
These surveys, interpretations mainly based on data of the Konjunkturinstitut and the
Statistische Reichsamt, may be recommended to the reader as an introduction to German
material.
THE WORLD CRISIS AND AFTER 973
acted the effects of foreign devaluations only in part, and was not greatly
income generating, since the direct subsidy was financed by a levy on
industry.
More important for our subject are the efforts directed toward aut-
arky, under the influence of which values of imports, which had recovered
somewhat in the second half of 1932, soon receded again and then fluctu-
ated about a horizontal line until their rise during 1937. This policy,
while its effect on welfare was, of course, negative, was an important
factor in stimulating prosperity. A large part of the new investments in
industry was for the development of resources that were to replace
imported materials, for example, for the development of the iron-ore or
the aluminum supply, of the production of synthetic gasoline or of the
staple fiber or of synthetic rubber (Buna). These instances illustrate
our distinction between conditioning and carrying out innovation. Had
the response to that policy consisted merely in an expansion on existing
lines, in an extension of, say, flax or wool production, this would have
been "passive adaptation" and that policy would have constituted com-
plete explanation. But that was not all. New things were done involving
the distinct entrepreneurial act that constitutes "creative adapta-
tion." In offering an opportunity for entrepreneurial activity in pre-
cisely those fields, that policy became no doubt responsible for many of
the innovations that carried the prosperity which may be dated from the
first quarter of 1935, but only in the sense implied in offering an oppor-
tunity (or providing a condition) with which an infinite number of
responses would have been compatible,1 In appraising effects it is,
however, not only necessary to take account of the damage done to other
parts of the industrial organism — which it is difficult to do — but also to
remember that it does not follow that there would have been less entre-
preneurial activity in the absence of the encouragement to press forward
in this particular direction. Entrepreneurs might simply have pressed
forward in others. The reader should observe that the failure to see this
simple truth or, as we may also put it, the habit of exalting the importance
of the particular opportunity that has actually been exploited, at the
expense of the role of the opportunity-exploiting force or agent, is an old
1 By stating that the policy of autarky, as such, conditioned but did not more than
condition a certain type of innovation, we do not mean that the government did not do
more than that. It gave leads. It exerted pressure. It helped in various ways in financ-
ing and promoting. Production of synthetic gasoline was, for example, subsidized by a
levy on the brown-coal industry. For the expansion of iron-ore production, the Reichs-
werke Hermann Goring were founded, of which the Reich took most of the common stock.
And there were many cases of pure state enterprise. This active leadership was, of course,
something very different from mere "control" or "regulation" and also from mere condi-
tioning. But it must be distinguished from the policy of autarky as such.
974 BUSINESS CYCLES
error which we have met over and over again on our way (for example, in
the discussions (Chap. VI) about the "rise of capitalism") and which is
a fertile source of faulty diagnosis.
In a different form, the same error must be guarded against in a
discussion of the conspicuous success of the spending policies of the Ger-
man government. We must strictly distinguish between two types: one
of them was simply of the pump-priming kind — the German words
being, as before, Ankurbelung and Arbeitsbeschaffung — in the sense that
it, and other expenditure induced by it, was additive to that system
expenditure which would have obtained in its absence, while the other
was substitutive in the sense that it took the place of part of that system
expenditure. This distinction does not coincide with that between
income-generating expenditure and non-income-generating expenditure,
for public expenditure may be strongly income generating and yet need
not be additive in that sense: it would not, if income-generating private
investment were, for example, at the same time curtailed by the same
amount. We cannot, of course, draw the line according to the objects
for which given income-generating outlays were made. Roads, canals,
public buildings, beautification of cities, armaments may all come into
either category, although we will call the second category "armaments/*
a potiori, in spite of the fact that such things as the Rhine-Danube Canal
(act of May 11, 1938) also enter into it. But we can pretty definitely
draw a line in time : income-generating expenditure, whatever its motive,
was primarily pump priming or additive until about the first quarter of
1935 and primarily substitutive after that.
Pump-priming income generation was, though only an enlarged edi-
tion of previous policies, part of the so-called First Four-year Plan and
of a recovery policy which shows much family likeness to that of the
United States. What might be termed the German AAA met a different
situation in a different spirit, but is similar to its American counterpart in
its aim of raising agricultural prices. The German NRA was the act of
July 15, 1933 (introducing compulsory cartels). It stressed restriction
of production and of real investment — by order of the Reichswirtschafts-
ministerium — still more than its American counterpart. Founding
new firms, constructing new plants and resuming operation of plants
that had been shut down, installing new machinery or otherwise increas-
ing capacity were for definite periods prohibited in a long list of industries.
Later on, these cartels were in other respects tightened and, as far as not
superseded by other controlling agencies — the commissioner for the super-
vision of pricing in particular; decree of Nov. 5, 1934, which revived an
office already created Dec. 8, 1931 — made a permanent instrument of
government control over production. But as a part of recovery policy,
prohibition to invest, except in special cases (e.g., prohibition of opening
THE WORLD CRISIS AND AFTER 975
new retail shops), was discontinued as soon as it was thought that it
had served its recovery purpose and replaced by compulsion to invest
according to government order. The first case (October 1934) thus
taken in hand was that of the brown-coal industry mentioned in the last
footnote. It was organized into a compulsory "community" (Gemein-
schaft) and directed to make a levy on its members with which to finance
the Braunkohle-Benzin A.G. for the purpose of developing hydrogenation.1
Within this arsenal of recovery measures pump priming went on by
means of relief and public works expenditure, as elsewhere largely
financed by short government paper discounted at the Reichsbank or
taken under rediscount promise of the Reichsbank by banks, saving
banks, and other institutions. It is not easy to estimate its amount, but
excluding the investment done by the state railways and the post office
the writer believes it to have been, until the middle of 1935, of the order of
magnitude of from 3 to 4 billion marks.2 Previous argument justifies
the statement that this expenditure, coming as did the American spending
program of 1933, after the event, i.e., after the lower turning point of
the cycle, did not break the spell of a depression that would otherwise
have kept the system indefinitely at the minimum point. But its suc-
cess in helping and accelerating recovery was striking, as shown by the
unbroken improvement and the rapid rate of absorption of unemploy-
ment. Since such success does not attend all pump priming, it is reason-
able to attribute it to the manner in which it was done in this case and to
concomitant policies. The sums disbursed were comparatively moder-
ate. They were expended with strict economy. Creation of purchasing
power was an incident, but it was not pursued as an end. Speculation
was not encouraged. Infraction of social discipline was discouraged.
No attempt was made to raise costs. Monetary wage rates, in particular,
were regulated with a view to stability at a level not much above depres-
sion minimum (see below). Saving and accumulation were encouraged,
or as little discouraged as possible, and in many instances enforced.3
1 As we shall see below, compulsion to invest in some lines frequently implied prohibi-
tion to invest in other lines; but these prohibitions were no longer dictated by the recovery
purpose and carry a different meaning. There was a third class, viz., the restrictions
dictated by the raw-material and exchange situation. These, however, do not interest
us here.
2 The three programs for providing employment (Arbeitsbeschaffungsprogramme)
alone amounted to 1,900 million marks. The writer has added the whole of the government
subsidies for repairing buildings (500 millions) and a highly conjectural part of the expendi-
ture on motor roads and other items of public construction not included in those 1,900
millions. The chief difficulty is in determining how much of the total created additional
income.
3 Space forbids our entering into technique, but the main headings must be mentioned.
First, that policy of wages was on balance conducive to increasing the sum total of savings.
Second, taxation was at first somewhat readjusted in a way favoring savings and later on,
976 BUSINESS CYCLES
And all this minimized dislocating and maximized precisely that part of
stimulating effects which does not produce relapses. The strength of the
" fascist" state as against group interests, and its fundamental attitude
to economic life — which for it is not an end in itself but a subordinate
servant of extra-economic national goals — in this case facilitated a
behavior in accordance with the rules of long-run economic rationality.
Government expenditure was not less conspicuous and government
leadership was more so in the subsequent prosperity than it had been
during the recovery. But it thenceforth acquired a different economic
character. We speak of armament expenditure as belonging to a differ-
ent category, neither because it was evidently not motivated as a pump-
priming measure — that would not matter — nor because it was not
income generating — for it was — but because it was not additive in the
sense defined above. There is, of course, no reason to believe that
results would have been any different if the government's demand had
been for washing machines and baby carriages instead of for war equip-
ment. But there is also no reason — other than an ad hoc assumption
based on prejudice — for believing that, once started upon the road toward
full employment, the system would have failed to reach it and to embark
upon a prosperity phase if there had been no such government demand at
all. For, on the one hand, an equal amount of existing monetary means
and of facilities for creating additional ones would have been released to
firms and households, and, on the other hand, private investment pro-
grams abounded and households were obviously not averse to expanding
private consumption. That this was so is proved by the fact that, in
order to finance, and to provide the physical resources for, its own pro-
gram of investment and consumption — guns are consumers* goods — the
government had to place severe restrictions on private issues, real invest-
ment, and consumption alike. In particular, the restrictions on issues
and investments other than those within the government program can
no longer have been motivated by a wish to overcome depression. Hence,
when the burden had to be increased (e.g. by the increase in the corporation tax from 20 to
30 per cent, August 1986), it was done in a way which did at least not differentiate against
saving. Third, by an act regulating the distribution of profits of corporate enterprise
(Dec. 4, 1934), limitations were imposed not on undivided but on distributed profits.
Fourth, where real investment in industrial plant and equipment was made compulsory,
this was done in such a way as to make saving or accumulation practically compulsory too.
Fifth, interest, while regulated, was regulated at a comparatively high level. This may
have had something to do with the increase in saving deposits — though the reader knows
how little faith the writer places in any figures about saving — which (annual increase in
saving banks and other institutions) was as follows: 620 million marks in 1933; 957 in 1934;
1,244 in 1935; 1,084 in 1936; 1,823 in 1937. The Reichsamt estimates (this, however,
is still more doubtful) the total amount of monetary savings at 1,571 millions for 1933 and
at 6,895 for 1937.
THE WORLD CRISIS AND AFTER 977
they also, like government expenditure, acquired a new meaning: they
were continued in order to reserve, for the government, factors of pro-
duction which were known to be on their way to other employments and
had to be deflected from them. But this implies that government did
not create a demand for those factors which would otherwise have been
lacking, but that it only substituted its own demand for that which would
otherwise have been forthcoming from other sources — which is what we
wished to prove.1
The argument does not deny that the presence of so obvious, strong,
and steady a demand greatly facilitated matters, and still less that the
government's labor and saving policies substantially contributed to the
success revealed by the output and employment figures. Nor does it
deny the importance of the momentous change involved in the fact of
actual management of industry by the state. But it still follows that
the prosperity was only state-directed and not state-created and that it,
hence, fits very much more closely into our schema than one would
believe at first sight. The developments to 1937, in fact, make perfectly
good Juglar recovery and prosperity phases. The interesting question as
to what inference that suggests about future cyclical phases in the Ger-
man economy and, in particular, depressive ones is perhaps not difficult
to answer. Theoretically it is possible so to plan the sequence of innova-
tions as to iron out cycles; but after supernormally strenuous periods of
advance there will be recessions in our sense even in the corporative state ;
most of the symptoms of depression, however, need not occur at all or
can be made to disappear promptly by so powerful a central authority.
It remains to survey in the light of the above analysis the behavior
of a few additional series. Prices were strictly regulated and, owing to
the practically complete socialization of Germany's international eco-
nomic relations, divorced from foreign levels and deprived of any influ-
ence on international movements of commodities. Moreover, they were
also subject to all sorts of distortions resulting from the German situation.
But they did not on that account entirely fail to conform to the logic of
our model. This is due to the fact that regulation asp racticed not only
1 For the obviousness of that proof we are indebted to the government's determination
to do the thing in as uninflationary a way as possible. If instead of — practically — com-
mandeering the productive resources of the nation, it had bid for them in free markets
(which could have been done only by means of much additional credit creation), its demand
would have been additive, and it would have been more difficult to show that in real
terms our diagnosis would have held true even then. It should be observed that while
a much more general (and also more correct) theoretical argument could be produced in
support of the economic truth we strive to grasp, our proof makes use only of the particular
conditions of present-day Germany and, hence, cannot itself be generalized. Nor does it
prove that the intensity of prosperity under state leadership was exactly what it would
have been without it. It might have been less intense. It might have been more so.
978 BUSINESS CYCLES
caused but also avoided some of those deviations which under modern
conditions of pricing we are likely to meet. During the five years that
have elapsed since the government came into power, the index of whole-
sale prices rose (April 1933 to April 1938) from 64.8 (1928 = 100) to 75.4.
But this rise was due mainly to the prices of agricultural products and of
imported materials — rubber rose by over 600 per cent. Finished indus-
trial producers' goods fell somewhat, while finished industrial consumers'
goods rose by almost 25 per cent. The official index of cost of living rose
during the same time by 8.5 per cent.1 The picture thus reflects the
tendency appropriate to the Kondratieff phase if we take account of a
number of circumstances peculiar to Germany. It does not reveal any
intention of government to raise the price level as such. Excepting the
agrarian sector, there seems in fact to have been no tendency to use
increase in prices as a stimulant. On the contrary, efforts were made to
prevent the expansion of the circulating medium from having that effect.
Nevertheless, and in spite of the continuing — though perhaps decreasing
— losses incident to the policy of forcing exports, of increases in operating
costs due to increasing prices of raw materials or to the unsatisfactory
nature of substitutes or to working beyond optimum capacity, and of
increases in total cost due to the necessity of liberal allowances for
depreciation, industrial net earnings increased through all the years
under survey. But they did so at a decreasing rate suggestive of the
approach of recession. So did stock prices.
That price policy was made possible by the government's wage policy,
which up to the point of full employment produced the very results that
we would expect under conditions of perfect competition. According to
the official data, average annual hourly rates rose to only 76 pfennigs for
1937 (from the minimum of 70 in 1933) roughly as cost of living2 and thus
remained far below the maximum of 1929 (96 pfennigs) in monetary, and
somewhat below it (by about 3 per cent) in real terms. This policy of
1 One reason why cost of living did not increase more than that was the unrelenting
pressure that was exerted on retailers' margins. In food retailing net revenue of many
small shops is said to have fallen to as little as 1,000 marks, the equivalent in purchasing
power of roughly $800 to $400 per year. Of course, this was due not only to price policies
but also to various restrictions which limited physical volume of sales, increased difficulties
in purchasing, and to other causes. Fundamentally, there is hardly room for the independ-
ent retailer of the capitalist type in a community that extends regulation to consumption.
2 If welfare considerations were relevant to our subject, we should have to take account,
on the one hand, of the change in the taxation of wages (3.5 per cent to 1929, then 3 per cent
to 1982, 3.5 per cent for 1933, then 4 per cent and 4.5 per cent for 1936 and 1937; including
poll tax; data from Reichsamt, Institut fiir Konjunkturforschung and reports of Reichs-
kreditgesellschaf t) though not the contributions to social insurance, and, on the other hand,
of a number of benefits, holidays, cheap excursions, marriage loans, protection against
dismissal, and the like that are difficult to evaluate but of considerable importance.
THE WORLD CRISIS AND AFTER 979
making and keeping labor a cheap factor of production greatly helped to
increase the total income of the working class. The industrial wage bill
just about doubled in utter disregard of high-wage theories between
1933 and 1937 (1933: 5,921 millions; 1937: 11,900 millions; the Saar
country being, for the sake of comparability, excluded in both years).
An investigation of the Berlin Institute states that about 65 per cent of
this increase is attributable to increase in numbers employed; about
11.5 per cent to longer hours; about 10 per cent to promotion to better
paid jobs; and about 13.5 per cent to increase in rates. The sum total
of all wages and salaries increased only from 12.1 billions in 1933 to 20.9
billions in 1937 and total national income only from 46.6 to 68.5 billions
(1929: 75.9), as a consequence of a policy which, on principle, strove to
conserve rather than to reverse the downward revision of monetary
values effected by the depression. Real per capita income increased all
along and in 1936 and 1937 surpassed that of 1929. Though the results
of recovery and prosperity were to a great extent absorbed by public
consumption and investment, the consumption of the masses expanded
in the field of the commodities of modern life, household gadgets and the
like. It contracted in some departments, where it ran up against the
exigencies of autarky. But per capita figures lend only feeble support
to the widespread belief that there was all-round contraction in the
consumption of food. Comparison of figures for 1937 and 1929 shows
indeed, significant decrease in the cases of citrus fruit, beer, eggs, and
margarine; and an insignificant one in the case of wheat. But per capita
consumption was higher, e.g., in the cases of rye, meat, fish, butter, sugar,
and coffee.
The management of money and credit was facilitated and its effective-
ness greatly increased by that policy of wages, by the government's
attitude toward and perfect control over savings, and by being completely
entrusted to one very able man. But that part of it which interests us
here1 did not, either in the problems to be solved or in the methods of
1 The rest of it, viz., the management of the relations of the German to foreign monetary
and credit systems — complete control of transactions in foreign exchange and all move-
ments of commodities and balances was only the most obvious tool of that management —
also bears upon our subject, not only through the "insulation" of the German money
market which it achieved, but also through its influence upon the investment process.
But we cannot enter into it. It must suffice to state that this part of German monetary
policy was chalked out for government and Reichsbank by the debt situation of 1932 and
the methods, stand-still agreements and moratoria, by which it was then handled. The
facts that certain extra-economic policies of the National- Socialist government produced a
"flight of capital" and that foreign credits were both difficult to get and from the national
standpoint undesirable, of course intensified the difficulties of that situation and of the
problems of raw-material supply. Foreign devaluations and some repayments and
repurchases, all of which reduced the foreign debt from about 19 billion marks at the begin-
980 BUSINESS CYCLES
solving them, fundamentally differ from English or American manage-
ment. The German money market being almost completely insulated —
much more so than it could have been insulated by a mere abandonment
of the gold parity — the Reichsbank, while allowing its holdings of gold
and exchange to dwindle to practically nothing (a little more than 70
million marks before the absorption of the Austrian National Bank),
secured powers to embark upon open-market operations and to "cover"
her notes exclusively by domestic bills and certain types of fixed-interest
securities through the legislation announced Oct. 17, 1933. It also
secured more stringent control over other banks through the Banking
Act (Gesetz iiber Kreditwesen) of 1934. Thus armed, it immediately
set about expanding the volume of credit, but only within the require-
ments of pump-priming public expenditures. To the end of 1935 it
increased — together with its affiliate, the Gold Discount Bank — what
we may term central credit outstanding (very roughly corresponding to
federal reserve credit outstanding) by about 2.7 billion marks, which
was sufficient to unfreeze what in 1933 and 1934 there still was to unfreeze
in the industrial and the banking structure and to help the market to
absorb the various kinds of short government paper which financed that
expenditure (about 1.5 millions of tax refund bills — Steuergutscheine,
introduced by the Papen plan — several billions of bills for the financing
of employment — Arbeitsbeschaffungswechsel — other special bills — Sonder-
wechsel — and ordinary treasury bills, all of which produced a net increase
in all kinds of bills outstanding of between 5 and 6 billions, or about
50 per cent, from the last quarter of 1932 to the last quarter of 1935).
No great pressure was brought to bear on short rates. In fact, the
method of allowing that mass of short paper to go to the open market
relieved member banks of idle funds and idle facilities and thus of the
necessity to compete for other outlets. Short rates fell in the natural
course of things, but remained comparatively high throughout : bank rate
ning of 1933 to about 10 billions at the beginning of 1938, brought but little relief. Under
these circumstances a policy developed from what at first had been temporary emergency
measures. It may be described either as an attempt to secure some effects of devaluation
without devaluating or, more tellingly, as graduated devaluation, the rate varying from
about 40 per cent to zero. The state, completely controlling foreign economic relations
and therefore being in the position of a discriminating monopolist, discriminated by
means of a schedule of prices for marks graduated according to the use to be made of every
one of them and to the elasticity of the demand in each use: holiday making in Germany,
for example, is for the foreigner a highly substitutable commodity, hence the offer of a
particularly cheap traveling mark. It follows that it is not correct to say without qualifica-
tion that Germany has not devaluated at all, but that it is also not correct to say that the
official parity price of the mark is meaningless and nothing but a sham. If it were a sham,
the method of export subsidies, which is highly distasteful both to government and indus-
try, obviously need not have been resorted to.
THE WORLD CRISIS AND AFTER 981
stood at 4 per cent from 1933 to 1938, rate on prime bankers* acceptances
(Privatdiskont) fell only from 3.88 to 2.88, daily money rate from 5.11 to
2.63. This need not have been so. But recovery was financed at
relatively high rates in order to facilitate normalization, which was kept
steadily in view. Long rates were in some sectors reduced by government
action, but where they were not, especially in the bond market, adapta-
tion to the liquidity in the short market was allowed to proceed but
slowly. Prices of 6 per cent industrial bonds reached parity only in
1935 — they remained above it after 1936, when the rate was reduced to
5 per cent — and the price of 4.5 per cent governments — until the conver-
sion of Mar. 31, 1935, the rate was 6 per cent — did not approach it until
May 1938 (99.7).
"Expansive" short financing continued during the prosperity, but it
was increasingly replaced by government issues offered for public sub-
scription. These had been insignificant in 1933 and 1934, but amounted
to 1,636 millions in 1935; 2,670 in 1936; 3,150 in 1937; and to 1,934 in
the first quarter of 1938,1 when a program of normalization was
announced. After one more issue of the "expansive" type — not redis-
countable, however, at the Reichsbank, though eligible as collateral for
loans — financing by " special bills " was to be discontinued; the Reichsbank
was to help in absorbing the existing ones but no longer to rediscount them
or any new ones; the Reich was to finance from taxation or ordinary
loans, industry and trade from ordinary bank credit. In itself and
barring disturbance by political events, this program is perfectly sound
both in the sense that it is possible to carry it out and that there is no
reason why doing so should produce a normalization crisis. But very
heavy taxation is an essential element of it. In consequence of the
increase in incomes and in rates of taxes, the revenue of the Reich
increased by 7.3 billions between the fiscal years 1932-1933 and 1937-
1938, and there is, of course, the decline in expenditure on unemployment
relief amounting to nearly 2.5 billions. But the tax on corporate earnings
contributed over 1,400 millions to that increase, and the income tax —
exclusive of the tax on wages — nearly 1,700. Whether in the fascist state
the effects of this will in the long run differ from those we should have
to expect in other social conditions, remains to be seen.
The behavior of banking series reflects the processes of state-directed
and state-financed recovery and prosperity, and can be easily interpreted
from this standpoint.2 When foreign observers in the second quarter
1 Redemption of all kinds of public debts during 1935 to 1938 amounted to 1,750
millions according to the Reichskreditgesellschaft.
2 This, however, applies only to the most general contours. As we have seen before this,
German bank statistics are anything but easy to analyze in detail. Difficulties have,
among other things because of the multiplication of intermediate instances (Arbeitsbe-
982 BUSINESS CYCLES
of 1933 inferred from the continuing shrinkage of loans and deposits —
by deposits we mean the German Kreditoren and not deposits in the
German sense — that there was no genuine recovery, they not only over-
looked that foreign devaluations and repayments to foreign creditors
would automatically produce a shrinkage of balance-sheet items, but also
that the domestic processes of unfreezing and of repaying emergency
credits would naturally have the same effect. But German observers
who did not fail to urge this, still expected that expansion of member
bank credit would presently set in. Consequently they were disap-
pointed at what looked like absence of "secondary effects," the failure
of public expenditure to stimulate private investment, and so on.1 In
fact, the Debitoren (loans) of the five great banks even declined through-
out prosperity and were but little over 3 billion marks by the end of 1937,
after having been a little over 4 billions at the end of 1933 and about 6
billions at the end of 1929. 2 Their deposits including "saving" deposits,
over 10 billions in 1929, moved on a falling level through 1935 and slowly
recovered to but 6.8 by April 1938. Cash also declined.3 Bills, however,
mostly of course the "special" bills, more than doubled from the first
quarter of 1933 to the last quarter of 1937 and investments — owned
securities, eigene Wertpapiere — also increased considerably.
There is nothing to wonder at in this picture. Its features are the
consequence of government financing. Business did not go to banks
for what it received from government, hence business expansion did not
directly show in bank statistics except in money in circulation — which
amounted to 5,418 at the end of February 1933, and to 7,219 at the end
of February 1938, the last figure unaffected by the annexation of Austria
schqffungsbanlcen and other) and because of the lumping of "special" with other bills,
not decreased since 1932. We shall confine ourselves to the Reichsbank and the big Berlin
banks, but will take this opportunity to mention two additional points. First, we have
seen what plight the municipalities maneuvered themselves into by their liberal expendi-
ture during the twenties, which during the crisis issued into complete breakdown of munici-
pal finances. One of the first tasks of financial reconstruction was, therefore, the refinanc-
ing especially of their short debts which was done by the Gemeindeumschuldungsgesetz.
Second, in order to keep the figure of Reichsbank credit outstanding as low as possible, the
Golddiskontbank issued one-name paper (Solawechsel) to the banks in order currently to
absorb their liquid means and to use these for relieving the Reichsbank.
1 For some of these comments, as well as for Professor Bresciani-Turroni's interpreta-
tion, see the latter's article, The Multiplier in Practice: Some Results of Recent German
Experience, Review of Economic Statistics, May 1938.
2 The commodity loans (Warenvorschusse) , which at the end of 1929 had reached 2 bil-
lions, were but a few hundred millions throughout 1935 to 1937. We do not emphasize the
decline in financing of international trade (Rembours-Kredit) and in stock exchange loans
(Reports) because they are accounted for by obvious special reasons.
3 Since a very large part of the other assets could at a moment's notice be liquidated at
the Reichsbank, there was, in fact, little object in holding cash beyond current requirements.
THE WORLD CRISIS AND AFTER 983
— and in the bills.1 And multiplier effects and the stimulation of private
investment show during recovery in the sphere of production and employ-
ment rather than in the sphere of regular bank credit, while there is no
point at all in looking for the former after full employment had been
reached. Those effects are, no doubt, difficult to evaluate, but during
what we have defined as the pump-priming period their presence seems
clear enough wherever they were not obstructed by difficulties about raw
materials and so on. During what we have called the armament period
— the prosperity phase — the monetary processes of an ordinary prosperity
were replaced by the monetary processes of a state-financed one, as other
processes were by the processes of "investment" for the purposes of
armament and of autarky (the Second Four-year Plan). The tapering
off of pump priming thus becomes invisible, because the flow of prosperity
expenditure also originated with the state, but it was nonetheless real.
Demand for bank loans was not forthcoming, because expansion in lines
other than those which government approved and largely financed was
impossible. For the same reason — which here stands out still more
clearly because of the prohibition of all but government-approved issues
— the late and weak revival in new industrial bond issues does not prove
anything. They were insignificant until 1937, when they amounted to
258 millions. Issues of shares increased steadily from the minimum in
1933, but amounted to only 395 in 1936 (333 in 1937; the maximum was
1438 in 1927). But no inference about absence or weakness of the entre-
preneurial impulse follows under the circumstances.
F. Recovery and Recovery Policy in the United States from 1933 to
1935. — What, according to our schema, should have been a Juglar
recovery covers the period from the autumn of 1932 to the first months
of 1935.2 Separate treatment is necessary, not only because of the
difference in phase, but because of the presence of another difference
which while, strictly speaking, but one of degree is so important as to
amount to one in kind: while, as we have seen, recovery policy was a
distinctly minor factor in 1931 and not of decisive though of greater
importance in 1932, it thenceforth dominates the scene. This is so obvi-
ous as to raise the question whether there is any sense at all in going on
speaking of cyclical phases and trying to date them, or to relate actual
1 The Reichsbank's holding of bills other than ordinary treasury bills was, for the same
two dates, 2,439 and 5,637 millions.
2 More precisely, our experimental count gives from the middle of November 1932
to March 1935 inclusive, and within this period a Kitchin recession to the end of August
1933, a Kitchin depression to the middle of June 1934, and a Kitchin revival covering the
remaining months. Again, this is not intended even as a reference schema but merely as
an illustration. But it should be observed that, properly understood, it is not out of step
with actual events.
984 BUSINESS CYCLES
business situations to our process. Many economists would not hesitate
to answer in the negative. Theories have, in fact, been offered which
are explicitly or implicitly based on the hypothesis that either from 1914
or from 1929 — the beginning according to some writers of a series of
completely new vicissitudes of capitalism that are unheard of in the
sense that previous history does not present anything at all comparable —
or, finally, from 1933, a new economic pattern has more or less suddenly
emerged which superseded the previous one for good and calls for a new
analytic model and fundamentally new assumptions as to both data and
mechanisms, especially with respect to the investment process.
We are not concerned with the methodology underlying those theories
and with their intrinsic merits or demerits.1 We are face to face not
with a question of principle but with a question of fact. The only princi-
ple involved is the one which has been stressed throughout this book and
which rests on the certainty that the economic process of capitalist society
will eventually turn into something fundamentally different and on the
ever-present possibility that our process be temporarily blotted out by
the action of more powerful factors, such as, for instance, ruled events in
Germany 1914 to 1923. As to the facts, we not only know that all the
essential features of the postwar period up to the world crisis, but also
those of the world crisis itself answer perfectly to expectation from our
model, i.e., from past experience. Moreover, we know something else
that sweeps from our path what otherwise would, as we know from the
theoretical discussion in Chap. IV, be an extremely thorny problem :
we need not ask whether the system "could have" recovered without
political action stimulating it out of a state of prostration. For it did.2
1 The writer's opinion about them is contained in the argument of this book as a whole
and cannot be put into a few words. But it is useful to remember that that question
divides up into the questions of logical correctness of those models and of their "fit," which
are entirely independent of each other. A pattern of reality, its meaning, and its tendencies
may be correctly seen by an economist who yet blunders in constructing his theoretic
model. And a model may be entirely correct in itself and yet fail to fit the facts and
especially the tendencies enshrined in contemporaneous fact. The present writer feels
that if he were at this moment to attempt the task of comprehensive criticism of current
theories, he would have to stress what seems to him a large amount of error of both kinds,
partly resting on unwarranted generalization from recent experience. But he also suspects
that if he were to attempt this task 10 years hence, he would have to defend the kernel of
truth that in both those respects these theories contain. Comments on some of those
theories will be offered at the end of Sec. G.
2 The author, who so often is painfully aware of the fact that his argument has to con-
tend against a powerful aversion to its real or supposed implications, thoroughly enjoys the
psychological vantage ground over which he is traveling at this point. For any aversion
of readers against accepting the writer's opinion will be much mitigated by the only alterna-
tive open to them, which to many would be no less distasteful: whoever refuses to believe
that the recovery in 1932 occurred in the ordinary course of the working of the system will
have to believe instead, that Mr. Hoover turned the tide. That his administration would,
THE WORLD CRISIS AND AFTER 985
This being so, affirmative answer to our question is unavoidable:
there is not only point in going on to relate the course of events to our
process — in the sense that we assume every one of the successive situa-
tions to have been the resultant of the working of this process and of the
effects of government action, both, of course, not only superimposing
themselves on but also influencing each other — but there is no choice
but to do so. For it would be contrary to all experience and com-
mon sense, though of course no logical impossibility, that a process,
which can be strictly proved to have been running its course since at
least the sixteenth century and right to the end of 1932, should have
come to a stop suddenly on Mar. 4, 1933. It should be observed, how-
ever, how severely our task restricts the scope of our discussion of the
policies of the period. What are their most important aspects to many —
to those in particular, who welcome them as the dawn of an epoch of social
reconstruction — must, if we are to focus our attention on the mere effects
of those policies on the process which is the subject of this book, be
excluded to the point of exposing the writer to the indictment that he is
completely lacking in social vision and in the grasp of the broad social
issues involved. While nothing can be done about this,1 it is hoped that
enough has been done to protect the argument that is to follow from the
different, though cognate, misunderstanding which may result from our
speaking of recovery policy as an external factor acting on our process.
It has been pointed out not only that economic — or any — policy grows out
of and is, though not uniquely, shaped by the economic situation with
which it attempts to deal, but also that in the case before us the short-run
situation in the spring of 1933 was such as to force any but the strongest
but for the whim of the political calendar, have come out with flying colors and amidst a
universal clapping of hands is, for reasons of political psychology, very plausible in any case.
1 The very fact that we refer to the economic policy of the period as recovery policy
seems to suggest not only narrowness of outlook but also the application of an altogether
unfair standard. Therefore, it should at least be explicitly stated that it is fully recog-
nized that those who framed or defended those policies aimed at much more than mere
recovery and, hence, had to strike a compromise between different and in many cases
conflicting aims. This conflict — "reform versus recovery" — was and is unavoidable and
does not in itself constitute a ground for the charge of logical inconsistency or of error in the
choice of means. But it introduced an element of prejudice and insincerity into the discus-
sion of the recovery problem. Since what a majority of people really wanted was recovery,
opponents of reforms developed a tendency to attribute to all such measures effects prejudi-
cial to it, and advocates of social reforms, a habit of wholesale denial of the presence of
that conflict. And it is against theories framed for these purposes that we have to protest.
Moreover, in many cases the element of personal or group interest was as obvious in the
arguments on the one side as in those on the other. In still others, psychotechnics —
for instance, attacking an uncomfortable piece of reasoning by impugning the motives or
sympathies really or supposedly behind it — were more in evidence than professional
competence.
986 BUSINESS CYCLES
hands1 irrespectively of prevailing preferences for or aversions to
"planning."
1. Thus narrowed down, our task may be further simplified by exclud-
ing a number of measures which cannot have had major effects on succes-
sive business situations or which cannot have had them before 1935.
Nobody will, for example, attribute major consequences to the Federal
Economy Act of Mar. 20, 1933, or to the revision of veterans' claims (Sec.
20 of Independent Offices Appropriation Act of June X6, 1933) or —
though some stimulating effect, extending beyond the industries directly
affected, is of course beyond doubt in this case — to the modification and
the subsequent repeal of prohibition. Most of the more than 80 acts
passed by the Seventy-third Congress up to June 16, 1933, may for our
purpose be dismissed as both uncontroversial and unimportant, although
the sum total of them, no doubt, influenced — mainly steadied— the exist-
ing situation. The only measure of this class for which this may be and
actually has been called into question is the Securities Act of May 27,
1933. The writer would have passed it by, thinking that it was not
only the kind of thing that has, ever since South Sea Bubbles days, often
been done after abnormally severe breakdowns revealing reckless financial
practice, but also a sober and well-drafted piece of legislation, from which
1 Two facts underlie what the reader will, according to his pleasure, call either the
criminal folly or the profound wisdom of American recovery policy. First, to a greater
extent than is generally admitted in popular discussion, the measures taken in and after
1933 but continue and develop what had been done or begun before. Second, the measures
of social reform actually taken run on familiar European lines, some inaugurated by men as
little radical as Bismarck and Taaffe (Count Edward Taaffe, Austrian prime minister
1880 to 1892), others outlined in the German Arbeitsrecht of the twenties. It seems bad
sociology to call the United States "backward" in this respect, but many of the most
characteristic New Deal measures would naturally be suggested by the simple consideration
that this is no longer a pioneer country. Two more remarks on New Deal policy : a measure
does not stand or fall with the arguments that are used by its promoters and cannot, for
example, be called foolish merely because those arguments are; and the ultimate aims a
measure is to serve are not only irrelevant for our purpose but extremely difficult, if not
impossible, to know. The monetary legislation of 1933 may illustrate this. It has —
even by some of its supporters — been called inflationary. We may perhaps feel fairly
confident in diagnosing the aims of the various "inflationist" groups — debtors, speculators,
silver men, exporters, and even, so it seems to the writer, banks which while standing for
financial decorum were presumably not all of them sorry to have their credits unfrozen —
but it is certainly beyond the writer to define unequivocally the subjective meaning of the
actual measures enacted. Given the fact, if it was the fact, that "inflationist" interests
were at the moment strong enough to override a veto and to have their way, the method
of having "inflationist" powers conferred on the president and thus gaining time for
recovery to gather force and for dealing with each of those groups separately might con-
ceivably have commended itself to a sound-money man of classical tradition, and no doubt,
bears interpretation in an anti-inflationist sense, especially as those groups were subse-
quently, in fact, separately satisfied in ways that salvaged considerable portions of that
classical tradition. This will be shown presently.
THE WORLD CRISIS AND AFTER 987
no depressive influence could emanate. But it caused a flutter, and not
only in interested quarters, the main point attacked being the construc-
tion of the liabilities imposed on issuing houses and security dealers.
These, however, do not seem to amount to more than responsibility for
what one "knows or ought to know" or to enforce more than the care
habitually taken by any decent firm. Much more plausible reasons than
this act are available in order to explain the stagnation in nonpublic
issues at that time.1
Recovery was substantially facilitated by the Emergency Banking
Act of Mar. 9, 1933, already mentioned, which provided machinery for
the reopening of closed banks,2 by the Banking Act of June 16, 1933,
which introduced a number of important reforms — the most important
refer to strengthening the Federal Reserve System's power over members,
particularly with a view to regulating extension of credit for speculative
purposes; to holding company and security affiliates; to stricter centraliza-
tion of open-market operations; to branch banking, and, for us most
important of all,3 to deposit "insurance" — and by Title II of the Emer-
gency Farm Relief Act of May 12, 1933, which dealt with the agricultural
credit and especially with the refinancing problem. The Farm Mortgage
Corporation Act of Jan. 31, 1934, and the Home Owners' Loan Acts of
June 13, 1933, and April 27, 1934, were similarly to cope with another
thoroughly frozen part of the credit structure and thus also to relieve
the banking situation — they look quite conservative with the homage
they pay to "local thrift." We will add further examples of measures
1 It has been pointed out to the writer that his interpretation rests on the hypothesis
that the provisions of the act are to be administered reasonably and without vindictiveness
or desire to victimize unpopular interests. That is so. It is also true that ill-founded suits
brought under the act may impair the position of a firm, even if its action be eventually
found unexceptionable. But whatever may be thought about these and similar arguments,
the conclusion that the act was no major impediment to recovery seems still to stand. For
an authoritative exposition of the opposite view, see C. J. Bullock, The Securities Act of
1938, Review of Economic Statistics, Jan. 15, 1934. If long-run effects were pertinent to
our present purpose, the case for that reform would be still stronger, however.
2 It has been mentioned before that this act and the amendment, Sec. 136, of the Federal
Reserve Act (Industrial Advances Act, June 19, 1934) also provided for powers for the
Federal reserve banks to lend to individuals, especially industrial concerns, but that these
powers were not used to a significant extent.
3 That Act represents the most systematic attempt that has been made to deal with
what, from the standpoint of the structure and technique of banking, always struck
observers as the most obvious shortcomings of the American banking system. This larger
aspect is beyond our sphere. But it should be mentioned that the problem of the small
and inefficient bank was attacked only indirectly and that the problem of long-term and
illiquid credit was not attacked at all, except by making such loans still easier. Anything
suggestive of restriction of credit, the speculative purpose alone excepted, was inacceptable
to public, legislators, and experts.
988 BUSINESS CYCLES
of this type: the extension in various directions (insurance companies
especially) of the scope of the Reconstruction Finance Corporation's
assistance by the Acts of June 10 and 14, 1933; the Emergency Railroad
Transportation Act providing facilities for consolidations, rationaliza-
tions, and reorganizations; and the United States Employment Service
Act of June 6, 1933. *
These and similar measures did not make recovery. They helped to
provide conditions, institutional and other, rather than stimuli, for the
process of recovery to resume quickly after the catastrophe of the spring
of 1933, by solving individual problems, removing impediments and
potential storm centers, constructing safeguards, allaying fears, and, on
balance, improving the general atmosphere — all of which would other-
wise have been the source of prolonged difficulties and waste. In
these respects the combined effect of the recovery measures of this type
must be rated high, although, if they had stood alone, we should have
no hesitation in speaking of a process of recovery propelled by the forces
embodied in our model.
2. But we shall arrive at no very different conclusion as regards the
two towering monuments of early New Deal policy, the AAA and the
NRA. All the great and small questions of principle that surround
them — especially those pertaining to the realm of welfare economics2 —
and all the ultimate effects they had or eventually would have had on the
institutional framework of the economic process of this country and on
this economic process itself being excluded from our discussion, we easily
arrive — discarding for the moment the provisions on money inserted
into the act creating the one and the labor provisions contained in the
other act — at the result that, on balance, both of them promoted recovery
of the usual type without replacing it by a recovery that would have to
be explained on different principles. They certainly paralyzed, and
replaced by others, certain parts of the ordinary capitalist machine but,
taking the national organism as a whole, in a way and to an extent which
was corrective rather than constructive.
1 The Farm Credit Act (June 16, 1988) was a measure of agricultural credit policy in
general, and has but little direct bearing on the particular emergency, since the safeguards
provided for severely limited the granting of distress loans. The Tennessee Valley Author-
ity Act of May 1933 would for us, but for its possible bearing upon private power finance
and investment, simply come in under the heading of government expenditure. The far-
reaching aspects with which public discussion has invested it and, as it were, the symbolic
significance it has thereby acquired are not relevant to our subject. See Professor E. S.
Mason on the Power Aspects of the TVA's Program, Quarterly Journal of Economics, vol.
50, p. 377.
2 Both cases, in fact, illustrate well the rationale of our distinction between welfare and
prosperity.
THE WOKLD CRISIS AND AFTER 989
This is especially clear in the case of the agricultural adjustment
policy.1 As we know from previous discussions, it was to deal, unlike
NRA, not with an emergency simply but with an emergency in which a
long development creating fundamentally untenable conditions had
suddenly come to a head. Wholesale liquidation of farms, impinging
on an industrial unemployment which itself was unmanageable for the
moment and which had set into motion a current of remigration to the
land, would have been the "automatic" method of restoring equilibrium.
The alternative to this — what we have called "orderly retreat" — pre-
cisely implied temporary or even — for those who from any of the many
reasons one may have for this desired to keep alive a large and contented
farming population — permanent preservation of disequilibrium in the
agrarian sector to be unavoidably financed by the (normal) surplus of
the industrial sector.2 To do this would, ipso facto, facilitate general
recovery. It would reestablish something like the previous processes
in the agrarian sector and the previous relations of the agrarian to the
industrial sector. Thus it would also relieve the debt and banking
situation and thereby stop up a source of actual and possible cumulative
depressive effects. At the same time it could not, the relative financial
strength of the two sectors being what it is, exert pressure on the non-
agricultural sector severe enough to open up another such source instead.
This argument applies a fortiori in case the means transferred were to be
created ad hoc or taken from sources other than that part of nonagrarian
incomes which was being currently spent. But it should be observed
that it would also apply if nonagrarian consumers' expenditure had been
reduced by an equal amount, i.e., if the agrarian Paul had really only
received consumers' dollars taken from the industrial Peter. The prob-
lem thus being perfectly clear and soluble, a very simple program suggests
itself of refinancing bona fide farmers threatened by foreclosure, of
nationalizing the marketing of, and particularly the export trade in,
agricultural products, and of strictly planning production — into which
program measures of more fundamental adjustment and of further
rationalization could have been inserted at will. Most of this being
out of the question under the Constitution and owing to the unprepared-
1 Since it is impossible to do justice to the subject, reference should be made to the
treatment of the recovery aspects of AAA in a series of publications of the Brookings
Institution. See, in particular, J. S. Davis, Wheat and the AAA, 1935; H. J. Richards,
Cotton and the AAA, 1936; J. D. Black, The Dairy Industry and AAA, 1935; H. B. Rowe,
Tobacco and the AAA, 1935; E. G. Nourse, Marketing Agreements under the AAA, 1935;
and especially Nourse, Davis, and Black, Three Years of the Agricultural Adjustment
Administration, 1937, particularly Chap. XIV.
2 From this standpoint, much talk about justice, income or price parities, in itself open
to rather obvious objections, acquires perfectly sound meaning.
990 BUSINESS CYCLES
ness of public opinion, the well-known devious route was chosen (Title
I of the Emergency Farm Relief Act of May 12, 1933), which raises a long
string of problems peculiar to it, incidence and other effects of processing
taxes among them.
But provided we agree that the net effect was to increase farm revenue
considerably, we need not go into those problems since then the rest, i.e.,
the proposition that a contribution was made to general recovery,
automatically follows. In order to take this view, it is not necessary for
us either to accept palpable exaggerations of the role that* the farmers'
plight played in the general depression, such as a prominent authority
has been guilty of in stating that 60 per cent of all the unemployed "lost
their jobs because of the reductions in rural buying power"; or to share
the opinion of fervent advocates about the effects of agricultural recovery
on general recovery, some of whom went so far as to call the advantages
accruing to agriculture a mere "incident" of a general benefit conferred
on the nation; or to fall back upon doubtful theories about effects through
an incident decrease in savings or, somewhat less incorrectly, nonspend-
ing; or, finally, to overlook the role of the droughts, of the depreciation
of the dollar, and of general recovery itself. Net results, which are all
we have, are at best difficult to interpret, and the panegyrics of the
administration of the agricultural adjustment act on its own activities
no doubt bear discounting. If, for example, we read in its first report
that in rural communities delinquent taxes were being paid, debts owing
to banks discharged, schools reopened, orders placed for clothing, fur-
nishings, implements, automobiles and parts, all because of AAA, we
cannot help feeling that a sound case is being spoiled by overstatement.
It is, moreover, not easy to determine how much of the results actually
attained is attributable to the basic idea, restriction of production —
"paying for not producing" — and how much to other devices that were
not, or not necessarily, bound up with it, such as marketing agreements
and semimonopolistic export practices. The energetic tobacco program
supported by the Kerr-Smith Act scored the most striking success, in
fact, considerably beyond the goal envisaged,1 owing to particularly
favorable conditions of demand. In dairying where, because of a
relatively sound fundamental situation, there was much less need for
action — leaders accepted what was offered to them rather than pressed
for it, and there was more opposition to than support for a produc-
tion program — restriction (purchase of dairy cattle) was secondary,
and market agreements — in some cases supplemented by what amounted
to internal import prohibitions — mainly did, not always successfully,
1 That goal itself was not defined as in the other cases. When it turned out that the
base period 1910 to 1914 yielded a "parity** which was already attained at existing prices,
reasons were found for a higher "fair exchange value."
THE WORLD CRISIS AND AFTER 991
what there was to be done.1 Leaving aside the somewhat equivocal
results of the corn-hog action and other items, we will recall that the
effect of the wheat program, whatever it might have been, was largely
absorbed by the dominant effects of four successive bad harvests — 1933
to 1936 — so that, disregarding various supplementary measures of minor
importance, the actual benefit to wheat farmers from AAA substantially
reduced — except perhaps for 19342 — to the benefit payments financed
by the processing tax which for the three years during which the arrange-
ment was in force, amounted to $326 million and may be likened to a
simple subsidy.
The cotton program was but to a minor degree interfered with by
adverse natural conditions. Participation was extensive from the outset
and became still more so under the pressure of the Bankhead Act and of
various privileges (seed loans, loans of the Commodity Credit Corpora-
tion) that were confined to participants. This actually did for cotton
what nature did for wheat. Curtailment of acreage for 1934 and 1935
was made more effective by destruction of roughly £5 per cent of the
1933 crop. (Annual) prices to growers for 1934 were nearly double the
(annual) prices for 1932 — which were depressed, however, by the govern-
ment holdings of over 3 million bales resulting from the operations of the
lSee the appraisal of J. D. Black in The Dairy Industry and AAA, 1935.
2 Weather and restriction of acreage — more precisely voluntary restriction — were, of
course, both "external factors impinging on the process of recovery" and acted in part
alternatively and in part cumulatively. But it is held that even in the latter case weather
was the more important factor. This may be and has been questioned for 1934, exponents
of the administration claiming that AAA accounted in that year for a reduction of the crop
by between 50 and 60 million bushels, or about 10 per cent. If this be so, then the state-
ment in the text would have to be modified for that year, 10 per cent being enough to make
a very considerable difference to price and value. Professor Davis (Wheat and the AAA,
p. 349) estimated the difference made by AAA to acreage sown at only about 5.4 per cent.
But even this would be far from negligible. It should be added that it would be nothing
for the administration to take pride in if they had intensified the effects of the drought,
nor a criticism of the AAA policy if they had failed to do so, but that on the contrary
the latter ought to be recorded to the credit of both the arrangement and its handling
by the administration. The writer believes that (again neglecting effects of supple-
mentary measures about exports and surplus clearing in the Northwest) the drought
and possibly, for 1934, AAA restriction, together with general recovery, substantially
account for the behavior of wheat prices, and that Dr. Davis (op, cit., p. 366) overestimates
the role of monetary policy. The striking rise that occurred from May to July 1933 was
certainly due to a speculative movement induced by inflationary talk, as well as by antici-
pation of a bad harvest and of further government action for the benefit of agriculture.
But the subsequent failure of prices to fall to, say, the 1931 average cannot, except to a
minor extent, be ascribed to monetary causes, because of the inadequate importance of
the connecting link, exports. We do not deny that monetary policy played some role and
shall mention it again below. But it is going too far to attribute to it "something like one-
half" of the increase from 1932 to 1933-1934.
992 BUSINESS CYCLES
Federal Farm Board 1929 to 1931, to the point of reconquering foreign
markets lost (partly) through the price-pegging policy of those years —
and revenue of growers rose from 483 millions to (inclusive of benefit
payments) 880 millions in 1933 and 893 millions in 1934. This was not
even all. For without the program, the crop of 1933 would have been
one of the largest on record. In this case, of course, monetary policy
also counted for more (see below) owing to the relative importance of the
export interest.1
Net costs and losses to the Treasury (including reduction of the import
duty on sugar) were about 900 millions — about one-third of this being
due to the invalidation of the processing tax — which may roughly be
said to constitute additional expenditure within the system. Apart from
this, such additional expenditure was involved in the running of the
scheme though it was originally intended to be self-supporting. What-
ever we may think about technique, details, aims professed, or arguments
used, the success of the policy in removing a major obstacle from the
road of recovery and in reviving shriveled tissues in the economic organ-
ism is beyond reasonable doubt.
3. Title I of the National Industrial Recovery Act of June 16, 1933,
as embodied in the codes of fair competition, introduced a type of state-
supervised industrial self-government the gist of which, stripped of
phraseological mimicry and apart from provisions about labor, was legal
recognition and official encouragement, amounting to compulsion, of a
modified form of the German cartel which, quite independently of this
legislation, tended to grow out of the activities of trade associations.
Thus it is not easy to understand — except on the supposition that the
mere word Planning and the mere sight of a government office are to some
people irresistibly attractive — the enthusiasm with which some "liberal"
economists greeted a measure which associated Planning with that very
restriction and price rigidity that are usually debited, as its greatest
blemishes, to the account of Big Business. But it is all the easier to
understand how it helped recovery: exactly as the German cartel,
it pegged weak spots within industries, stopped spirals in many places,
mended disorganized markets, especially in cases of inelastic demand and
of that "overproduction" which is incident to the process of underselling
1 The Cotton Option Plan of the first year and the 10-cent and 12-cent loan policy had
effects on the international position of the United States cotton to some extent similar to
those of the Farm Board's pegging policy, which preceded it. On this and other points
see Professor Black's article, The Outlook for American Cotton, Review of Economic Sta-
tistics, Mar. 15, 1985. The present writer, being in any case more pessimistic than Profes-
sor Black about the long-run possibilities of American cotton exports, does not rate very
highly the unfavorable effects which policies of the type discussed may, in the long run,
have on them. We are, be it repeated, merely speaking of immediate effects during that
recovery.
THE WORLD CRISIS AND AFTER 993
the obsolete. There is little doubt about its effectiveness in paralyzing,
in some instances, the process of industrial transformation that was going
on: the failure to see that there was such a process at all or, in fact,
anything else but breakdown and deadlock was part of the philosophy
of the time. There is as little doubt about its effectiveness in improving,
in other instances, situations in which lack of organization really wrought
wanton destruction as in the cases of oil and bituminous coal1 and others
to which a less sweeping and spectacular measure could have confined
itself.
Immediate results for the general business situation were, however,
only the stronger because of the range over which this policy was indis-
criminately applied — although very unequally enforced — and so were its
purely psychological effects, which in a situation of that kind we have a
right to consider an important factor — even Blue Eagles do count for
something when, objective conditions for revival being given, it is broken
morale that is the matter. Invalidation by the Supreme Court (June
1935) came when the end had been achieved and was for the administra-
tion a blessing in disguise. But aftereffects were not entirely eliminated
thereby. Business had learned a lesson. The "chiseler" continued to
be frowned upon. And we shall have to bear in mind that there is here
a possible, if only partial, explanation of the fact that output figures
failed to come up to expectation in the subsequent prosperity phase,
which precisely for this reason most economists preferred to call imperfect
recovery.2
1 Section 9 of the Act dealt specifically with oil restriction. If it did not do so in the
manner of the Kaligesetz, this was presumably only due to lack of powers and the necessity
of avoiding constitutional pitfalls. In the bituminous coal industry technological advance,
increasingly economical use of coal, and increasing use of other fuels had created, ever since
the war, conditions which afford a class-room example both for the necessity and the diffi-
culty of regulation. In 1933 and 1934 price control was tried, but it lapsed with the
Supreme Court's invalidating the Bituminous Coal Conservation Act (1935).
2 The Report of the President's Committee of Industrial Analysis formed to prepare a
"review of the effects of the administration of Title I of the National Industrial Recovery
Act" after the latter had been declared unconstitutional, is not in itself very interesting.
But the five studies of the committee's staff (unpublished but open to consultation) and
many official and nonofficial studies — especially those that emanated from the Research
and Planning Division — making use of materials collected under the administration of the
Act, are not only an important source of historical material but also a storehouse of prob-
lems and patterns for the analytical economist quite inadequately exploited so far. The
general literature of the subject of NRA suffers both from insufficient foundation in the
details of actual fact and from the preconceptions of authors, which in the majority of
cases are obvious from the start and even affect statements of fact. A well-chosen sample
of the more significant publications has been ably discussed by Professor Rogin in The New
Deal, Quarterly Journal of Economics, February 1935. But such study as the present
writer has been able to make has convinced him that for the very limited purpose in hand
the view expressed in the above paragraphs would be substantially accepted by a majority
994 BUSINESS CYCLES
The quid pro quo which the Act and the Codes offered to possible
opposition from the ranks of labor and to possible criticism of what might
easily have been called antisocial tendencies, were the labor provisions.
Many an objector was reconciled by the clauses on child labor, hours, and
minimum wages which all codes (including the blanket code) contained,
or gratified by the protection extended in Sec. 7a to collective bargaining,
to organizing activity, and so on, which went far beyond the Norris-
LaGuardia Anti-Injunction Act of 1932. The larger aspects of the
progress thereby achieved in social legislation being once more beyond
our scope — of a progress that many of us will consider to have been
overdue — our only question concerns the effect on recovery of general
labor policy and of wage policy in particular. As to the first, the present
writer is confident that no inhibiting effect can be proved for the period
under consideration.1 As to the second, it follows from previous argu-
ment that, under the conditions of this country and of the prevailing
cyclical phrase, the persistent official efforts to raise the whole structure
of wage rates must on balance have had an adverse effect both on the
expansion of output and on employment per unit of output. While this
effect was probably small during the first great upward rush in 1933, the
further development of output, pay rolls, and especially employment,
which can hardly be said to come up to expectation, substantiates the
presence of this brake (see below, sub 6). The reader will realize that
this is perfectly compatible with an opposite result in many individual
cases; with such truth as there is in antisaving arguments; and with
recognition of the facts that the rise in price level partly absorbed and
government spending partly counteracted2 that effect. Again, and
of the economists who worked in that field, barring perhaps the analogy with the German
cartel and the last sentence. The reader is referred to Professor E. S. Mason's admirable
contribution on Controlling Industry in The Economics of the Recovery Program (D. V.
Brown and others, 1984), and to L. S. Lyon and others, The National Recovery Administra-
tion: an Analysis and Appraisal, 1935.
1 In other words, the mere recognition of the right of collective bargaining and the
elimination of the yellow-dog contract cannot in themselves have worked against recovery.
It does not follow that an organized drive, facilitated by the newly acquired opportunities
for action, or further legislation on the same lines might not have done so. But no such
drive or legislation occurred during the two years during which the NRA regulations of
wages and hours, paralleled by regulations by other federal agencies and state govern-
ments, were in force. The question will hence have to be touched upon again in the next
section.
2 Since government spending was in part a function of the amount of unemployment,
it can even be said that any unemployment created (relatively to what unemployment
there would otherwise have been) by that wage policy may possibly have induced a net
increase of total expenditure in the system. It will be seen, however, that this does not
invalidate our argument. The Research and Planning Division of the NRA is responsible
for a valuable source book (mim., January 1935) on Hours, Wages and Employment under
the Codes.
THE WOKLD CBISIS AND AFTEH 995
notwithstanding the rise in price level, labor was made expensive rela-
tively to real capital. Coupled with a cheap money policy, a high-wage-
rate policy was, under the circumstances of phase and country, the very
recipe for the production of a maximum of unemployment. It should be
observed, however, that while dampening recovery, this would not
necessarily affect the general character and the duration of the phase.
Let us pause for a moment in order to take stock. So far we have
before us the following elements of the situation in 1933. There was
first incipient recovery dating from about the middle of 1932. This
incipient recovery, second, had been interrupted by the banking
catastrophe in the spring of 1933 and was weighed down, independently
of that, by the conditions in the agrarian sector, by the results and
remains of a preceding state of overindebtedness in general, and by
depressive factors special to certain individual industries. Third, on this
pattern impinged the series of measures mentioned, all of which, with the
exception of the high-wage-rate policy inaugurated — though this was of
small immediate importance in 1933- — were on balance remedial in effect,
i.e., not only devised in order to remove those millstones but actually
effective in achieving this. Nothing more than these three groups of
facts is necessary for us to expect a strong and even violent rebound of the
system,1 more than compensating for the subnormal revival during the
preceding six months, to be interrupted, however, by the setting in of a
Kitchin depression late in the year.
This may be expressed by means of terms which have been so uncriti-
cally used as to elicit, when used now, little else but contempt: natural
and sound recovery. To the former term we assign the meaning of
recovery coming about in the course of the cyclical process by virtue of
its mechanism. The latter term we define as a recovery that is brought
about by factors which do not carry an inherent tendency to reproduce the
same difficulties as, or to produce other difficulties in place of, those that
they have been instrumental in overcoming: relapse is the most obvious
instance of the first type, and undoing of such work of readjustment as
may result from the processes of depression is an instance of the second.
It will be seen that natural and sound recovery are not made synonymous.
Whether "natural" recovery is always "sound" depends on whether we
exclude all those "understandable nonessentials " which may easily land
the system in an untenable situation. But inasmuch as depression itself
1 That statement, forming as it does an important link in our diagnostic argument,
should perhaps be amplified and buttressed. But the writer does not see how this could
be done without intolerable repetition. It should, however, be obvious that stopping
spirals and reversing "psychology" would, in a situation fundamentally ready for improve-
ment and with levels only un9erstandable from spirals and psychology, produce a strong
spurt.
996 BUSINESS CYCLES
is a pathological process, sound recovery need not be natural. In our
case the recovery we now envisage would — whatever might have been
the ultimate effects of NRA — have been substantially sound. And a
natural recovery was at the bottom of it. But the midwife role of public
authority was so important that it would not do to draw laissez-faire
conclusions.
It is not implied that Congress or the administration "should" or
"could" have stopped at the measures so far mentioned. The prevail-
ing unemployment would in itself suffice to refute any such implication.
But it is implied as a matter of diagnosis that to a greater extent than is
commonly realized the recovery of 1933 can be accounted for irrespec-
tively of the monetary and spending policies inaugurated at that time.
For this proposition there is, it is true, no such proof as the constellation
of facts enabled us to give for the cognate proposition that recovery
started "of itself" and not simply in response to recovery policy. But
there is a prima-facie case which is much strengthened by the fact that
during the critical second and third quarters of 1933 the new spending
program, being still in the incubating stage, cannot have had major, at
least mechanical, effects1 on the economic process. That substantially
the same holds true of the monetary policy will be evident from the
cursory discussion2 to which we now turn.
4. When banks reopened after the "holiday" in March 1933, they
were estopped from paying out gold except by special Treasury per-
mission— redemption of notes in gold was discontinued on Mar. 4 — but
only an insignificant fall in the international value of the dollar occurred.
On the afternoon of Apr. 19, i.e., immediately after the declaration of the
(partial) embargo on gold, dollars were internationally traded in at a
discount of about 9 per cent and during the following eight days at a dis-
count which fluctuated between 8 and 12 per cent. There was no panic
such as had occurred in the English case. This is highly significant. It
shows, on the one hand, that international speculation did not believe in
a substantial depreciation of the dollar — in spite of the fact that a bill
1 According to Mr. Currie's and Mr. Viilard's figures, published by Mr. Arthur D.
Gayer (op. cit.), net federal income-generating expenditure for the year was 1,856 millions,
only 210 millions more than in 1982. If the effect was more than proportionately greater
than this, it can only have been due to the difference in cyclical phase.
2 Strictly confined to the limits set by our purpose, that discussion not only neglects
details and technical aspects — such as the sequestration of gold, the abrogation of the gold
clause, the stabilization fund, later on the tripartite agreement, and so on — but also majoi
problems which for that purpose are of minor importance. The reader is referred to the
literature on the subject, but Professor Harris's Exchange Depreciation, 1936, and also
his article on the American Experience, Quarterly Journal of Economics, August 1934, should
be specifically mentioned. We neglect, in particular, the silver policy although, owing to
the nonnegligible importance of the relations with silver countries, some effects on cyclical
situations have doubtless been exerted by it.
THE WORLD CRISIS AND AFTER 997
was introduced on Apr. 20 which contained the main provisions that were
later embodied in the Thomas amendment — and on the other hand, that
in the absence of such speculative anticipation there was no reason
why the dollar, if left to itself, should fall at all. There was, thus, a
complete absence of parallelism between the fate of the dollar and the
fate of the English pound. The dollar was not under economic pressure
either in a short-run or in a long-run sense, either from abroad or at
home. We recall in particular that it had weathered the preceding
storm, the third through which it had had to sail during the depression:
the reserve system had the situation well in hand, and in less than three
weeks after the bank holiday over half of the amount was repaid which
had previously been borrowed from reserve banks in order to meet loss of
gold and withdrawals of notes from all centers and of bankers' balances
from New York. And almost all the losses of reserves suffered by reserve
banks were made good.
What pressure there was, was entirely political, and it was not until
the world became convinced of the imminence of "inflation" that the
dollar really gave way. This conviction was the result of the passing of
Title III of the Agricultural Adjustment Act of May 12, which was even
officially referred to as Emergency Relief and Inflation Act. We are not
going to waste space on an otiose discussion about whether the course
actually followed during the period deserves to be called inflationary or
not, and if so, in what sense. Although there cannot be any doubt that
the interests, the exponents of which swapped votes in order to insure the
passage of that act, aimed at inflation in every conceivable sense of the
term, the act itself was the result of a compromise that yielded less ground
to them than it seemed to yield and than, from a wide variety of motives,
almost everyone pretended to believe. Its provisions, besides being
only enabling and not mandatory, offered plenty of opportunity, sub-
sequently extended,1 to defeat any kind of inflation at will, and effective
use was made of it up to 1937. It was clear enough, however, that at
1 Section 46 of the act, amending Sec. 19 of the Federal Reserve Act, gave power to
the Federal Reserve Board to change the amount of reserve balances which members are
required to keep against deposits. The Banking Act of 1938 prohibits members to act as
agents for corporations and individuals in making loans on securities (loans on account of
others). The Securities Exchange Act of June 6, 1934, gives the Board of Governors power
to prescribe margins for brokers' loans to their customers and for security loans by banks.
The Gold Reserve Act of Jan. 30, 1934, has been so constructed as partly to eliminate the
"inflationary" effect of devaluation. The Treasury's braking power, already great, has
then been strengthened by the Social Security legislation which gives it control over vast
funds that may be used for the purpose. These are only some of the brakes built into the
machinery. To be sure, there are limits to their effects, and most of them must be put into
operation by visible and unpopular action. They therefore do not do away with the
possibility of inflation in every sense of the term. But it should be obvious how risky
numerical predictions about the future behavior of the price level must be.
998 BUSINESS CYCLES
least devaluation would be unavoidable. Even so, the dollar manifested
its natural strength by the hesitancy with which it fell. Therefore,
when in the autumn recovery dimmed and NRA and AAA enthusiasms
cooled, when, moreover, the announcement by the Reconstruction
Finance Corporation of its willingness to lend to banks up to a billion for
the purpose of relending failed, as it naturally would, to produce results,
the administration encountered the consequent inflationist onslaught by
resorting in October to a method which would bring $own the dollar
without "inflation," viz.9 the gold-buying policy,1 and by speedily
investing the political capital thus gained in putting a stop, for the
time being at least, to this type of "experiment" by the Gold Reserve
(Devaluation) Act.2 And still the dollar resisted: a torrent of gold
turned toward this country. We will but glance at the movements in its
gold stock during the February following upon the presidential proclama-
tion of Jan. 31, which raised the price of the fine ounce from $20.67 to $35
and the value of the monetary gold stock — including gold previously
acquired by the Reconstruction Finance Corporation and the Treasury,
but excluding coin still reported as in circulation — to 7.03 billions. No
less than 381 millions — a record — were imported during the month (213
of which from England) chiefly in response to the new price, banks at
home and abroad taking advantage of the undervaluation of the dollar.
There was also some release (8.6 millions) from earmark.
In order to appraise the effects of this policy on the economic process,
it is first of all necessary to realize how much or how little it had to do with
easy money. There is no doubt, of course, that the influx of gold, which
1 Inflation of the German type would, of course, automatically send down the dollar.
It was to association with such inflation (though not necessarily with exactly that dose of
it) that part of the support of the policy of depreciation and devaluation was due. What
this group of "inflationists" overlooked was the possibility of having the token without the
substance. Although facilitating future inflation, reduction of the gold value of the dollar
by the gold-buying policy — direct action on exchanges by offering unlimited amounts of
dollar balances would have been a still more effective method — was really (whatever the
intention) a means of avoiding it. Its tactical virtue consisted in the fact that, in spite of
being a device to avoid inflation, it would satisfy certain inflationary interests, e.g., export-
ers and speculators, who were particularly vocal and supplied the motive power of much
of the inflationary propaganda. Hence, it would break the inflationary phalanx.
2 It will be observed that our interpretation runs in terms of tactics and interests rather
than in terms of "theories" and "experiments," and severely disregards phraseology.
This, so it seems to the writer, is as it should be if grasp of situations and not phraseology
be our aim. Nothing is implied about actual motivation. But that interpretation has
until 1937 been quite consistently, borne out by facts. Although inflationary phrases,
"theories," and ends were freely made to serve tactical purposes later on, in each case
pressure on the administration was more obvious than pressure by the administration.
The incessant, though mistaken, appeals to banks to lend freely acquire their significance
in this light. For utilization of existing facilities would draw the ground from under the
feet of advocates of "currency inflation."
THE WORLD CRISIS AND AFTER 999
already in February 1934 carried excess reserves of all member banks to
the new peak of above 1 billion, was then and later the chief factor
responsible for what, at least from the spring of 1935 on, will strike us as
abnormally low rates. With government expenditure what it was and
reviving business, for example prime bankers' acceptances (90 days)
could hardly have reached % of 1 per cent by November 1934 and stayed
there, nor prime commercial paper % to 1 per cent by June and fallen to
% after that. But if such lows were beyond expectation from our model,
prevalence of very low rates was not, and it is not very obvious that, if,
for instance, New York City customers* rates steadily fell from their
modest "panic peak" of 4.88 for March 1933, to 2.64 for March 1935,
this was substantially more than we should have expected to find without
the gold movements induced by devaluation. In other words, the latter
did not create the conditions of monetary ease. The commercial paper
rate was down to 2 per cent by the autumn of 1931 and, in spite of the
subsequent rise, to less than that by the autumn of 1932, and there is
nothing in the processes of incipient recovery to enforce an upturn. As
far as this goes, devaluation did not lift any weight from the economic
process, as is indeed obvious from the behavior above described of the
dollar in 1933. Moreover, such pressure on rates as there was during
that year was due not only to the depreciation but also to the open-
market purchases of the Federal reserve banks, which in response to the
greenback threat contained in the Inflation Act acquired 570 millions
of governments from May to November and then stopped because this
only served to swell excess reserves. But that step would have been
possible without going off gold. Its ineffectiveness finally — a last veri-
fication of our views on the subject — goes far toward establishing the
proposition that, whatever influence on rates and credit facilities was
exerted by whatever external factor, the influence of these on the economic
process was next to nothing.
Devaluation must, in the second place, be considered in relation to the
policy of public expenditure. It has, in fact, been held that the meaning
of the former primarily consists in its implementing the latter, which was
what really produced results. There is, of course, some truth in this
view, which is at any rate much superior to the naive belief that redefining
the gold content of the dollar would per se change the price level in the
same proportion — a curious survival from the days of the commodity
theory of money. For although increase in price level neither is, as a
matter of principle, nor has been, as a matter of fact in this instance, the
main effect of antidepression public expenditure, it is true that devalua-
tion can, with the qualification to be mentioned presently, only act on
prices if and when it either induces or facilitates increase in expenditure.
Therefore, if public expenditure that would not otherwise have been
1000 BUSINESS CYCLES
technically possible had been made so by devaluation, we should have
to list the latter among the major factors influencing the recovery process.
But public expenditure perfectly adequate to produce the results that
actually were produced would have been possible with the dollar at the
old gold par. Devaluation may have facilitated it by removing all con-
cern about monetary limitations, but this is all.
There remain, in the third place, direct effects. We will mention
two. The year 1933 was one of monetary disorder ancj of widespread
apprehensions about impending inflation. An impulse was thereby
given to speculation in securities and commodities, which was very obvi-
ous on stock and produce exchanges. Stock prices (and prices of second-
class bonds), in particular, reacted visibly and until September con-
sistently to every drop of the international value of the dollar. The only
question is how far this effect extended beyond speculation in a wide
sense of the term. We have seen that the response of productive business
to monetary policy is, to say the least, equivocal. The notable instances
of 1896 and 1878 should suffice to show that anticipation of monetary
expansion is not necessarily a propelling, and anticipation of "sound"
money not necessarily a depressing influence. Nor is this at all astonish-
ing. Hence, although the anticipations induced by going off gold
certainly contributed to the hectic rise of the wholesale price index
— which, as has been emphasized by Mr. Snyder, reflects speculative
movements so disconcertingly well — which occurred from March to
July 1933, and although this spurt in turn no doubt had some effect on
productive operations, we shall not weight that component very highly,
especially in view of the fact that, as pointed out before, strong revival
would in any case be understandable without it.
The other immediate effect of depreciation and devaluation is on
foreign trade. The changes during our period in raw figures of either
values or quantities — which are not impressive — do not in themselves
prove much. As to export of United States merchandise, they are com-
patible with the opinion that the motorcar and machinery industries
benefited somewhat. Since, however, industrial exports are in any case
not important enough to matter greatly, we need not enter into the nice
questions surrounding any effort to isolate the influence of devaluation.
The official index of quantity of agricultural products exported continued
its downward course throughout the period, the annual figure for 1935
being a little less than 58 per cent of that for 1932. Some benefit to
wheat and especially to cotton farming is, nevertheless, beyond doubt.
In the latter case it also served to counteract the effect of the depreciation
of the rupee, the Egyptian pound, and the milreis. However, this
instance only strengthens the case for the broad proposition that, with
qualifications which need not be repeated, public spending was the only
THE WORLD CRISIS AND AFTER 1001
positively propelling measure acting on our process — as distinguished
from the measures previously discussed, which mainly removed obstacles
and thus (again, mainly1) helped in what may be termed a negative way.
5. Prima-faeie federal income-generating expenditure actually was
from about December 1933 to about the middle of 1937 the dominant
factor in the increase of net national income (current dollars2).
We will not go into the methods and agencies by which the spending
program was put into effect and which, along with the immediate objects
of the spending activity, were in an incessant process of change from the
time that foundations were laid by the Unemployment Relief Act of
Mar. 31, 1933 (forestation, prevention of soil erosion, plant disease, and
so on), the Emergency Relief Act of May 12, 1933 (revision of powers of
the Reconstruction Finance Corporation, creation of the Federal Relief
Administration, grants to states for the purpose of direct unemployment
relief) and titles II and III of the National Recovery Act of June 16
(Public Works Program, grants to states for highway construction,
appropriation of 3.3 billions). Methods and objects are not indifferent
either for the recovery or any other aspect. Even for the immediate
effect on the economic process it is by no means indifferent whether a
given sum be spent on direct relief or on purchase, from stock, of materials
the proceeds of which are used by recipients for repaying debts : there is,
according to the way of spending, a continuous variety of effects which
range from increasing system expenditure by several times the amount
spent down to not much more than refinancing.3 Nevertheless, we will
1 Some, especially AAA, also added to total expenditure in the system and as far as
they did, now come in again by virtue of this title. In a small way, of course, all of them
did so, owing to the administrative expenditure involved.
2 See Professor Kuznets' National Income, p. 8, Table 1, Col. 3. The figures for net
federal income-increasing expenditure are again Mr. Gayer's (op. cit.). It should be
added that Mr. Currie's method, on which they are based, includes also receipts from
estate taxes, on the ground that under the circumstances estates, even when able to pay
those taxes from owned cash and "near-cash," would do so from idle funds, which were
thereby put into circulation. It will be seen that income-increasing expenditure greatly
differs from the deficit as officially figured. All expenditure incident to refinancing trans-
actions is excluded, while income-generating expenditure of nominally independent agencies
(trust funds and so on) is included. Correction for variations in the income-generating
expenditure of states and local bodies would not make a significant difference.
3 Another point may be mentioned here. There is a type of public works which
amounts to little more than direct relief. But on the other end of the scale there is a
type which, even if undertaken from the relief motive, is nothing but a businesslike reaction
of public bodies to the prices of factors and the rates of interest which obtain in a period of
depression or incipient recovery. Strictly, expenditure of this character should be excluded
from an estimate of remedial expenditure, because it is not an external factor acting on the
process but an element of the latter's normal or autonomous mechanism. As state enter-
prise expands, this must become increasingly important, but it is so even with many
expenditures which do not issue in monetary returns at all or in returns which are com-
1002 BUSINESS CYCLES
merely note that net national income increased by about 8.6 billions in
1934 and by about 5.2 billions in 1935 (and by 8.81 billions in 1936), which
compares with federal net income-generating expenditure of 1,856
millions for 1933, of 3,238 millions in 1934, and of 3,154 millions in 1935
(and 4,025 millions in 1936).
As measured by those and other figures — for instance, of employment
(see below) — effects may well seem surprisingly small. They have, in
fact, been felt to be so even by those economists who simply attribute the
whole of the observed increase in national income to federal income
generation. But obviously we cannot do this, for to hold that income
generation alone has been responsible for that increase involves either
circular reasoning or else the theory that in the absence of it the
economic process would have gone on shrinking or would have dragged
along indefinitely at the minimum level.2 There is no warrant for believ-
ing this. On the contrary, there is, as we have seen, reason to believe
that there would have been recovery in any case — a recovery strong
enough to produce by itself most of the increases, especially in output,
that actually occurred and more. It follows that by taking account of
the cyclical phases on which federal income generation impinged, our
expectations as to its effects can be only raised and not lowered. We will
restate these expectations under four heads.
First, government expenditure will improve any business situation,
even if it increase the national income of the year only by the amount
spent or by less than that or even, in a limiting case, by nothing at all,
through helping the public to build up depleted balances and to repay
debts. If firms repay bank loans by means of money which the spending
government raised from existing but idle deposits, improvement may even
be accompanied by a decline in total or total demand deposits which,
as we have seen before, may thus be a favorable symptom.3 As a rule,
mercially adequate: public works using means of production that would otherwise go to
waste may be said to have been carried out at no social cost whatever.
1 Professor Kuznets' figures end with 1935. Hence, the figure for 1936, taken from the
Department of Commerce series, is not strictly comparable with the preceding ones.
2 An assumption of this kind underlies the reasoning of Professors Colm and Lehmann
in their otherwise excellent study of Public Spending and Recovery in the United States,
Social Research, May 1936. Dr. Currie (paper read at the meeting of the American Eco-
nomic Association, Dec. 30, 1937) argues that business deposits increased strongly from
1933 to 1935, while business borrowing declined, that this increase can only have been due
to the net income-increasing disbursements by government, and that hence public spend-
ing must have been the initiating force of recovery. This is a clear non sequitur, unless
we make recovery synonymous with increase in deposits and assume quite gratuitously
that the decline in borrowing was completely independent of those disbursments. See
the concluding paragraphs of this subsection.
8 Of course, that symptom is in this country not visible in any deposit series. The
behavior of loans, however, suggests that this component was not entirely absent.
THE WORLD CRISIS AND AFTER 1003
only a moderate amount of unfreezing is left for the early stages of the
recovery phase. But owing to the extent of the preceding catastrophe
and of the continuing state of overindebtedness, it is reasonable to assume
that in this case "consolidation" was one of the major remedial effects of
the spending policy.1 Though acting in the monetary sphere, it would,
however, primarily show outside of it.
Second, there were what we will call the direct results of handling
the government's money as far as it was not absorbed by the replenishing
of balances and the repayment of debts: the unemployed man spending
his dole, the man who has been reemployed in order to fill a government
order spending his wages, at the retailer's shop, the retailer thereupon
placing additional orders, and so on. Separate evaluation of this effect,
which was no doubt considerable, is impossible in the present state of our
information, since among other things we do not know the value of that
income efficiency of money (Chap. XI, Sec. A) by which the relevant
part of government-created income would have to be multiplied.2
Third, firms — particularly in recovery — will react not only directly
to government orders or to purchases by the first recipients of government
funds but also indirectly by expanding operations in anticipation of those
orders or purchases and by otherwise "magnifying" the immediate
effects of government disbursements. It should be observed, however,
that under the conditions prevailing in the early stages of recovery
new investment cannot be expected to be much in evidence. If in this
instance it had been, this would before 1935 have introduced an entirely
abnormal feature. This is not, of course, to deny that stimulation of
investment is entitled to a prominent place in a general theory of govern-
mental income generation. In and after 1935 some such stimulation
may have been present. In depression and recovery, however, it is
1 If we call that effect remedial — and the same term could be applied to the effects to be
noticed under the remaining heads — we do not thereby "justify" the spending policy.
A drug may be "remedial" with respect to headache but "injurious" to the heart. The
writer entertains no doubt but that public income generation outside of "deep" depression
(roughly from the middle of 1930 to the middle of 1931) impairs the efficiency of the
capitalist process for reasons that should be familiar by now. They apply especially to
public income generation during the later stages of recovery and the first stages of pros-
perity. The difference in the effects of public and entrepreneurial expenditure, the latter
involving, the former not involving, change in production functions, should particularly
be borne in mind.
2 We might, of course, figure out average income "velocity," i.e., efficiency times rate
of spending, in the usual way by dividing total income by deposits plus money in hand-to-
hand circulation. But it will be seen on reflection that we cannot arrive at the effect of
federal income-generating expenditure by multiplying the result into the amount of that
expenditure. Doing this would precisely involve the circular reasoning adverted to above.
It is not held that the problem is insoluble, but its solution involves a formidable research
program.
1004 BUSINESS CYCLES
current operations in and near the consumers' sphere that need to be
and, as a matter of fact, have been stimulated.
Fourth, irrespectively of reactions of the type noticed under the pre-
ceding heading, there will be ulterior effects on economic activity. The
relief in the debt structure, the steadying of prices, the improvement
in the sectors immediately affected by government disbursements, the
general feeling that a floor is being provided will remove inhibitions and
invite advance all round. This class of effects should have been particu-
larly strong in a situation in which not only the stage was set for recovery
but in which a recovery that had already begun had been interrupted
by an experience so trying to business nerves as an epidemic among
banks. While there is thus no intention on the part of the writer to
deny the reality of these effects, protest must be entered against the
practice of some economists in making an uncritical use of them, which
amounts to begging the question.
It should be added that federal income generation must also have
given an impulse to consumers' credit by making, directly and indirectly,
many households "credit worthy" which had previously ceased to be so.
1934, in fact, displays much higher figures than 1933 for the credits
outstanding of intermediary and cash-lending agencies and also for
the receivables of retail merchants. This was much facilitated by the
fact that this type of financing afforded at that time the most obvious
chance for banks to respond to the incessant appeals to "liberalize"
lending. Since the above analysis applies, with but little modification,
to income generation by consumers' credit as well as to income generation
by government expenditure, it is easy to realize that the former reinforced
the effects of the latter.
This analysis evidently harbors no tendency to underrate the poten-
tialities of pump priming which, as has been pointed out before and as
the war experience shows, may even if sufficiently persisted in turn
depression into a state displaying all surface characteristics of prosperity.
It will be observed that in some points we do not share the disappointment
felt alike by advocates of governmental income generation and by its
opponents. We were not, for instance, disappointed on the ground
that private investment was not more strongly stimulated in 1934. Nor
did we consider that application of government-created funds to the
replenishing of balances and to the repayment of debt constituted pro
tanto defeat of the spending policy. Observed results were, nevertheless,
no better than they could have been expected to be had that policy been
the only component to act. Since we are not at liberty to disregard
the other component, which also was adequate to produce those results,
we are driven to the conclusion — to be verified in the next section — that
other factors weakened the combined effects of both.
THE WORLD CRISIS AND AFTER 1005
Since so many economists accept the quantitative adequacy of the
injection of purchasing power to produce observed results as an ipso facto
proof that there cannot have been any contribution from the economic
process itself — or that there was a negative one — it will be well to retrace
our steps, in order to state explicitly the case for alternative possibilities.
One has been noticed under our first heading: that part of net income-
generating expenditure which increased national income only by its
own amount or less will, as long as we do not know its amount, obviously
be credited with more than its share if expansion unconnected with it
should occur at the same time — financed, let us say, from existing depos-
its— and if, on the strength of a plausible quantitative relation between
total income-generating expenditure and total increase in national
income, the one be connected with the other. The state of overindebted-
ness, on the one hand, and the cyclical phase on the other, combine to
make it a practical certainty that this coincidence was of some importance
in shaping the statistical picture.
Again, if government funds swell business deposits, as of course they
did, they will then finance the subsequent transactions of the recipient
firms, whether those transactions are induced by the act of expenditure
or not. It does not follow that every expansion of operations by these
firms must, therefore, be causally related to those receipts and that with-
out them they would not have expanded operations at all. In this case
it is indeed still more obvious than it is in the other how government
expenditure propelled recovery. But since resumption or expansion
is, for that cyclical juncture, independently motivated, it will not do
to attribute all that was financed by money originally inserted by
government disbursements to the impulse imparted by them: it can-
not be inferred that government carried business as Aeneas carried
Anchises. To some extent at least, financing by receipts simply replaced
financing by borrowing. As we shall see, there is ample evidence for this.
It should not be replied that this is a case of speculation about possi-
bilities vs. hard statistical fact. It is a case of common sense vs. a type
of monetary theories. What we see is the income-generating expenditure
and certain developments. The relation between them we do not see.
Our interpretation of it is not more but less hypothetical or speculative
than the one which exclusively relies on mechanical relations between the
two, because it assumes and asserts much less.
Finally, it should be observed that no account has been taken of the
possibility that government spending might have interfered with business
expansion (excepting, of course, the expansion of bank loans). Some
of the arguments adduced for this possibility fully merit the shrugging
of shoulders with which they are usually met, for instance the argument
that the unbalanced budget destroyed confidence. Others do not.
1006 BUSINESS CYCLES
But the net effect of the spending policy, taken by itself and considered
only with reference to the general complexion of short-run business situations^
seems to the writer to have been so clearly positive as to justify him in
disregarding, for the purpose in hand, any possibilities of that kind.1
6. We will now glance at the statistical picture. It should be recalled
that we need not lean too much on "inflationary" anticipations or on
speculation in general when explaining the boom of 1933, which culmi-
nated in July. Nor need we call it simply an affair of restocking (inven-
tory boom). Though all these and other factors — such as attempts to
profit from purchases or production not yet burdened by the extra costs
that were to be expected from AAA and NRA — no doubt materially
helped, that boom is fundamentally understandable as a belated and
hence more violent reaction to the ravages of the spiral, the intensity
of which it may reasonably be said to measure. Nor is it necessary
to emphasize the element of reaction to excess which undoubtedly
intensified the sharp slump — it was sharper than almost any that had
preceded — in the third, and the almost level movements of the fourth
quarter: a relapse was quite within ordinary regularity (Kitchin depres-
sion). But public spending no doubt shortened the relapse and accentu-
ated the upswing — which came in December — as well as the strong
expansion in the first half of 1934. There was but moderate relaxation
in the third quarter of 1934, and expansion resumed in the fourth
quarter. It went on at a much increased rate in the first quarter of 1935.
Outside debits (141 cities; Federal Reserve Board) express the funda-
mental contour very well from month to month. For the years in
question, monthly averages (which were at a maximum of 27.66 billions
in 1929; at 23 billions for 1930; at 12.87 billions in 1932) struck their
low point of 12.2 billions in 1933 (as we know, mainly in consequence of
the banking troubles) and rose to but 13.83 billions for 1934 (to 15.85
billions for 1935). In the absence of government income generation
we should have called this conforming to expectation. As it was, the
smallness of the increase becomes a problem. Our solution of this
problem is indicated in the preceding paragraphs.
Profits fell from the fourth quarter of 1932 to the first quarter of 1933,
if allowance be made for seasonal behavior,2 and recovered strongly in
the second quarter, but less than uncorrected figures suggest. Then
they relapsed again but made a showing much beyond expectation
1 C/., however, the last footnote but one. Neither NRA codes nor showers of money
are calculated to vitalize the capitalist process.
2 Cf. W. L. Crum, The Course of Corporation Profits, Review of Economic Statistics,
Mar. 15, 1934. The writer would, more positively than Professor Crum, assert that the
true low point of corporation profits comes in the second quarter in 1932. Improvement
for the greater part of the second half of that year is, in any case, beyond question, at least
as far as Professor Crum's selected list of 163 corporations is representative.
THE WORLD CRISIS AND AFTER 1007
from our model in 1934, the quarterly average for that year (Federal
Reserve Bank of New York) being about 75 per cent above that of 1933 —
roughly speaking, the writer would attribute three-quarters of this
increase to government spending and the NRA.1 Stock prices copy this
contour fairly closely. It should not be forgotten, however, that they
merely continued the upswing of the third quarter of 1932. The monthly
average of new corporate capital issues (domestic; Financial and Com-
mercial Chronicle; maximum 1929: 666.8 millions) touched its low of 13.8
millions in 1933 and increased to but 14.8 millions in 1934. This is even
less than we should have expected on the strength of the proposition
that recovery does not typically or necessarily start from real investment
and still less from capital issues. Government disbursements in part
supplied the funds which it would otherwise have been necessary to raise
by issues or to borrow from banks. Also, the proceeds of the issues of
1929 were still largely unused.
Some idea of the extent to which government disbursements replaced
bank credit by enabling firms to finance from receipts can be gathered
from the behavior of All Other Loans of reporting member banks. Those
loans did not increase, as we have seen, in the incipient recovery of 1932
— which conforms to expectation — but went on falling and ended their
downward course by a sharp drop in the first quarter of 1933. Then
they did rise moderately until the beginning of the fourth quarter, when
they decreased again. So far there was nothing abnormal. But they
did not at all participate in the upswing of the first half of 1934, while
United States securities held increased by about 1 billion in the first
quarter and by another half toward the end of the second. Increase in
the third was weak and short-lived and more than compensated by the
fall to the end of the year, and no increase accompanied the strong
upswing of the first quarter of 1935: firms did not go to banks for what
they got from government. It will be recalled that we observed the same
phenomenon in Germany.
Outside net demand deposits had increased during the second half
of 1932. After the understandable slump of the following spring — which
brought them to a low that is spurious in our sense — they increased more
than All Other Loans to July 1933, mainly, of course, because of the
increase in members' investment incident to federal reserve banks'
open-market purchases. Their rate of turnover also increased at the
same time. Then they fell and rose along with outside debits until
the middle of 1934, investment again taking the place of loans in
creating them. But they continued to increase after that, in response
1 Monthly averages of cash dividends (The New York Times; maximum 1930: $386.5
million; banks excluded) were 216.5 millions in 1932, 181.5 millions in 1933, 206 millions in
1934 (and 226.3 millions for 1935).
1008 BUSINESS CYCLES
to investment that did not take the place of loans which would otherwise
have emerged, until the end of the period: government by selling defici-
ency bills — the cash deficit of 1934 in the sense of the Daily Statements
was over 4.5 billions — acquired deposits which, when used, produced
other deposits: the old method of war finance. Transfer of surplus funds
to centers and influx of gold — at the new par — of course, also swelled net
demand deposits.1 There is no problem in any of these movements or
in the growth of excess reserves.
We have noticed how interest rates behaved under the monetary
regime that obtained, and will only add that bond yields (corporate
issues; the combined index of Standard Statistics) fell2 (monthly average)
from 6.27 per cent in 1932 to 5.92 per cent in 1933 and 4.86 per cent in
1934, and were 4.78 per cent for March 1935. The B.L.S. index of whole-
sale prices, excluding foods and farm products, rose from its minimum of
65.3 in April 1933, to not more than a monthly average of 78.4 for 19343
(75 for 1931), the greater part of the rise occurring before 1933 was
out, i.e., before the spending program had had time to produce its full
mechanical effects. This is within the limits of what might have been
expected as the result of a rebound from panic lows. Besides the
presence of underutilized resources, the weight of incessantly increased
productive efficiency is mainly responsible for the failure of prices to
respond more strongly to the price-raising policy4 of the administra-
tion. The latter was more "successful" with respect to the prices
of foods and farm products. Prices received by farmers increased
by 84 per cent (index of U. S. Department of Agriculture) between
March 1933 and December 1934. The B.L.S. index for farm products
rose, however, from 40.9 for February 1933, to 78.3 at the end of our
period, viz., for March 1935, i.e., by over 91 per cent of its minimum.
But the reader will realize that it would be easy to choose dates in
such a way as to get almost exact correspondence between that increase
and the decrease in the gold content of the dollar — he will also realize
how perfectly meaningless that would be.
1 Through August 23, 1935, the term net demand deposits retains its old meaning. Total
deposits of reporting member banks increased by about 3 billions during the year.
2 Along with a fall in stock prices, bond yields increased somewhat in the third quarter of
1934, while there was a temporary outflow of gold. This was a not unnatural concomitant
of deficit financiering, but on the whole such effects were successfully avoided. However,
bond yields were not low enough obviously to negative the idea of the presence of distrust
in the currency. The fact that really cheap rates were confined to the open market (in our
sense) does point in that direction.
3 The Fairchild retail price index, which also excludes foods (December 1930 = 100)
and also has a minimum (69.4) for April 1933, rose to a monthly average of 88.3 for 1934.
* Prices in full- weight dollars fell until November 1933 and did not really start to rise
until May 1934. But there is no need of going into the various aspects of this.
THE WORLD CRISIS AND AFTER 1009
Average hourly money wage rates behaved very differently in differ-
ent sectors — in anthracite coal mining they increased very little, in
bituminous coal very much, for instance, which again illustrates the lack
of realism incident to speaking of a wage level. On the whole, however,
these different rates of increase seem to have worked in the direction of
a more balanced wage structure. For manufacturing industries as a
whole ("all wage earners": 25 industries) an increase of over 22 per cent
occurred between June (the minimum) and December 1933. After
that there was no relapse — though there was one in the wages paid by the
wholesale and retail trade — but an increase at a much slower rate,
which was, nevertheless, sufficient to produce an increase in real rates of
about 6 per cent by March 1935. The annual average of the hourly
rates of skilled and semiskilled labor in manufacturing industries (National
Industrial Conference Board) was 55 cents for 1933 and a little over 64
in 1934. l This is clearly contrary to expectation from our model.
Excepting, perhaps, a fraction of the increase in 1933 which might have
resulted in any case from the general reaction to panic lows, it must be
attributed to government policies which are the only available factors to
explain how the upward shift of the demand " curve*' for labor could
have produced such rates in spite of the prevailing unemployment. No
comments about consequences need be added to what has been said on
the subject before (see especially above, sub 3).
Both factory pay rolls and employment started on their upward course
before monetary wage rates. In fact, it is important to recall that they
began to increase in the third quarter of 1932 and that the increase which
occurred from March to September 1933 was but a continuation and not a
break. Monthly averages (combined index of the Department of Labor)
of factory pay rolls were for 1933 about 6 per cent above 1932, for 1934
nearly 21 per cent above 1933, and for 1935 a little over 13 per cent above
1934, while the corresponding increases in employment averages were 10,
13, and a little over 4 per cent, the latter figure being suggestive of
inhibitions. Particulars of incidence of unemployment or reemployment
were as we should expect and need not be discussed here.
Output of manufacturing and mining, as measured by the Federal
Reserve Board's index and as reflected in production of electric power —
which, even if corrected for "trend," increased by 19 per cent from March
to July — and in carloadings, also behaved according to expectation from
our model modified by the political factor. Building contracts, for
example, were, of course, affected by the public works program and would
1 Per capita weekly earnings in manufacturing began to rise earlier — already in the
spring — and then increased with fluctuations to the third quarter of 1986. Of course, they
also increased in real terms, hours worked per week rising somewhat above their 1932
minimum of 34.8.
1010 BUSINESS CYCLES
not otherwise have increased much in 1933 — private building was indeed
below expectation, if anything. The sharp increase in steel-ingot produc-
tion in the second quarter of 1933 was from a very low level, but even so
surprisingly great — over 300 per cent in 4 months, thereby reaching the
1923 to 1925 average — so that the subsequent reaction of over 50 per cent
was very understandable. Neither its rise nor its fall, however, meant
what they used to mean of old, when steel satisfactorily represented
equipment. But it should be observed that from any reserve capacity
which may have existed in 1929 and which was obviously greatly increased
by the shrinkage of the subsequent years, a substantial deduction must
be made on the score of wear and tear — some of it independent of degree
of utilization — and of obsolescence.
Although machine tool orders rose to a fairly high level at the end of
the period, steel output was, during 1933 and 1934, primarily associated
with durable consumers' goods. Output of automobiles nearly reached
3 million units (passenger cars and trucks) in 1934.1 Refrigerators,
air-conditioning installations,2 and other members of the class of "depres-
sionless industries," such as gasoline, cigarettes, rayon, and some chemi-
cals, showed considerable gains.
In the behavior of output three points must be noticed. First, short
fluctuations do not quite correspond to our idea of what they should have
been. The increase of 66 per cent from March to July 1933, which carried
it to the 1923 to 1925 average and the subsequent decline are at least
timed according to expectation. But the increase in the first half of
1934 came 6 months earlier than we should have expected — relapse in
fact followed — and for the decline and stagnation in the first half of 1935
we have only political influences ( ?) to offer. Second, the index did not,
even before the decline early in 1935 and excepting the 1933 peak, reach
the level of 1925-1926, the preceding neighborhood of equilibrium, while
it should have surpassed it. We attribute this to the severity of the
depression and may recall that the recovery after 1873 also was not satis-
factory in every way. Nevertheless, effects of the NRA policy may be
partly responsible. Third, we must supplement the case for an adverse
effect of the increase in wage rates on employment by noting the striking
difference between the latter and the corresponding output. To that
66 per cent increase in output corresponds an increase of only 33 per cent
in factory employment. This is in part due to underutilization of the
working force employed around the bottom of the depression — average
1 This was not simply "rebound." Much had changed in the industry during depres-
sion years, and part of that increase must be attributed to its own impulse. There is a
kernel of truth in the exaggeration that "motors led us out of depression."
2 The latter, however, were not significant quantitatively. They are more properly
considered as an innovation of the following Juglar.
THE WORLD CRISIS AND AFTER
1011
hours worked per week, which may be taken as an indicator, in fact,
increased by nearly 5 per cent from 1932 to 1933 — and in part simply the
consequence of labor-saving rationalization which had been going on
through the depression. But it is impossible to overlook the premium
that the wage policy set on this rationalization.1
To sum up: The reader will have no difficulty in listing the symptoms
by which the effects of governmental income generation unmistakably
show. In spite of them, however, the statistical picture presented does
not differ fundamentally from what we should have expected to see in
the absence of that factor. Since, as has been pointed out in 5, it would
have been possible for such a picture to emerge, public expenditure not-
withstanding, under the sole influence of the normal recovery process,
the conclusion seems to suggest itself that — barring minor deviations
caused by it, such as we observed in the behavior of deposits — that
expenditure took no effect. This conclusion we do not draw. If we
did, we should, in fact, be committing the very same error which those
economists commit who simply attribute to government spending every-
thing that happened. But we do draw two other conclusions: first, that
attributing all observed developments to the normal recovery process
would, though wrong, not be more so — or more "speculative" — than
the opposite opinion; and that the prima-facie impression with which we
started in 5 is misleading.
G. The Disappointing Juglar. — If past experience be a guide and our
schema a roughly correct expression of it, then the rise of a new Juglar —
the fifth of the Kondratieff — was due for the spring of 1935, however
little meaning we may attach to the precise date yielded by our experi-
mental count (beginning of April). Though, as the reader knows and
as we shall presently sec, facts did not entirely fail, for a time just about
equal to the duration of an average Juglar prosperity, to bear out the
expectation which would follow from that, the difference is great indeed
CHANGES IN HOURS, EARNINGS, PRODUCTION, AND PRICES, MANUFACTURING INDUSTRIES, BY MONTHS
103£-1935
(Source: Monthly Labor Review)
Year,
monthly average
Index of
total
man-hours
Index of
average
hourly
earnings
Index of
production
Index of
average
output per
man-hour
Index of
labor cost
per unit
of output
Wholesale
prices for all
commodities
other than
farm products
1982
100 0
100 0
100 0
100.0
100 0
100 0
1933
107 3
98 1
119 0
111 0
88 0
101 0
1934
114 4
116 0
124 0
108 0
108 0
112.6
1935
126 2
120 1
143.0
118 0
107.0
117.4
1012 BUSINESS CYCLES
between such upswing as there was and what happened in the last
comparable instance, 1879-1880. Conditions external to this country
which then produced an agricultural boom may no doubt be invoked in
partial explanation. But government policy largely did for agriculture
now what European demand had done for it then, and if we take account
of government expenditure in general, agrarian and other, the picture
becomes still more disappointing. This was, in fact, universally felt.
People never spoke of more than recovery and an unsatisfactory one at
that. It would not, however, help us much if we did the* same; for the
real trouble with expectation from our model is not in the weakness of
that "prosperity" but in the fact that it was followed, instead of by a
recession in our sense, by a break which landed the system, within a few
months and at a rate surpassing everything witnessed during the years
from 1930 to 1932, in a state displaying all the phenomena of deep
depression. This would be still more unexpected — from the standpoint
of our model as such — as a sequel to a recovery than it is as a sequel to a
prosperity phase. Does it mean that the capitalist process has spent its
force, that private investment opportunity has vanished to the point
of making it dependent on government expenditure for motive power or
in such a way that the system must collapse as soon as government expen-
diture is withdrawn, like one of those children's balloons that shrivel as
soon as one ceases to blow air into them?
Reasons for believing that this is unlikely have been offered in the
introduction to the preceding section. But there seems to be evidence
for it of an almost experimental nature. The figures of federal income-
generating expenditure for 1935 and 1936 have been mentioned.1 Again
it was, prima facie, quite sufficient to justify the statement that it
induced, directly and indirectly, such — this time — prosperity as there
was during those two years, and if we take into account deferred effects,
also during the first half of 1937. Though there were developments that
seem to be beyond the range of its consequences, it might still be said
without absurdity that the spending policy then "took effect at last."
After that it was discontinued. On cash account the Treasury got out of
the red.2 A slump ensued in due course. Moreover, the writer enter-
tains no doubt not only that that slump will give way to recovery as the
new spending program within the 4 billion deficit budgeted for 1938
1 The veterans* cash bonus of 1,900 millions became payable from June 15, 1936. Half
of it enters into Mr. Currie's figures, being distributed, as seems reasonable, over the month
of payment, the two months preceding, and the three months following. The writer would
have distributed the whole of it in some such way. As a matter of fact, about 1,200 millions
were promptly cashed, and chain- and department-store sales responded immediately.
2 The figure of income-generating expenditure for 1987, $900 million, must be interpreted
with due regard to the fact that this sum was almost entirely spent in the first seven
months of the year.
THE WORLD CRISIS AND AFTER 1013
unfolds during the fall of 1938, but also that tapering off will again be
attended by the symptoms of — according to the way in which it is
effected — recession or depression.1 This should make us both envious
and thankful: envious because fellow economists will be able to enjoy
so delightful a verification of their views, thankful because in other fields
— medicine, for instance — people do not reason like that, or else we should
all of us be morphinists by now.
1. In order to see more precisely what there is to explain, we will
begin by a survey of time-series contours. 1935 was the third year to
show almost consistently higher levels of annual figures than its prede-
cessor. Monthly figures, however, were not consistently higher. The
weekly operating rate of the steel industry may serve as an example.
Excepting the first two months, it was at or below the 1934 level until
nearly the end of June; only in the second half of the year did it rise
above that. By end of September it was 50 per cent — in itself sufficient
proof of the weakness of that upswing, though 90 per cent was eventually
reached (March—April 1937). Moreover, there were other irregularities,
among them two setbacks in which leading series behaved in a somewhat
discordant manner: during the first half of the year it was the index of
production of manufactures and minerals and allied indices, such as
carloadings, that sagged, while outside debits steadily increased;2 the
little relapse in the autumn shows primarily in outside debits, while
production was hardly affected at all. These relapses and discordances
were — though on a very small scale — repeated in 1936, 3 which otherwise
1 That sentence has been left standing as it was written in July or August 1938. It may
be useful to add the following comments (May 1989).
Marked improvement in fact showed in the third and fourth quarters of 1938, espe-
cially in the indices of manufacturing output, construction, carloadings, department-store
sales, employment, though, conforming to expectation, prices continued to decline. These
facts obviously bear out the first part of the statement in our text.
But the relapse during the first quarter of 1939, which continued during April, does
not illustrate the second part. There was indeed, a considerable increase in Treasury
deposits during February, due to the sale of savings bonds and of securities issued on
behalf of various public credit agencies, which for a few weeks raised government cash
receipts above government disbursements. This, however, was hardly adequate to pro-
duce the observed results, and beyond this there was nothing but talk. We are still within
the rising tide of spending, and if effects do not show more visibly, this is due, apart from
the presence of depressing extrasystematic factors, to the cyclical phase. If the reader
refer to our schema he will see that, barring possible reactions to the abnormal slump of
1937-1938, there is little reason to expect from the mechanism of our process any very
strong upturn for several years to come.
2 That in the beginning of the prosperity phase money volume of transactions should
increase more than physical output is not in itself an irregularity; but output fell, especially
in steel, cotton and silk textiles, and bituminous coal.
3 But recalling our experimental count we shall not, as we must for 1935, consider the
occurrence of those setbacks as abnormal.
1014 BUSINESS CYCLES
is the year of the strongest and most nearly uninterrupted increase all
round. At its end the 1929 peak of the Federal Reserve Board's index
of production of manufacturing and mining was almost reached (with
120 per cent of the 1923 to 1925 average; from May 1935 to May 1936
the index increased by 16 points, and the value in current dollars of
national income produced (63.8 billions) was almost 80 per cent of the
1929 figure.1 Aggregate profits of 700 industrial and mercantile com-
panies (Federal reserve bank of New York), which had in 1935 been about
80 per cent above 1934, further increased by about 50 pe*r cent in 1936,
aviation heading the list and steel, automobiles, tires, petroleum,
chemicals and drugs, machinery and tools showing up particularly well.
Only 6.4 per cent of those companies reported net losses.
As stated above, a Juglar recession in our sense was due for 1937
(the middle of August in the experimental count). But what actually
happened was very different. After a drop in January, outside debits
recovered through May — not, however, to the figure of the preceding
December — then hovered on a horizontal level until August — so far con-
forming to expectation — and after that shrank rapidly. In the first
half of 1938 they continued to decline — not merely in function of falling
prices — but at a decreasing rate. By June they seemed to have reached
an even level. Output of manufacturing and mining behaved irregularly
from the first, and more irregularly than debits in the second half of
1937; instead of the increase that we should expect, we find that it
declined in January and recovered only to May, after which it fell by
about one-third until the middle of 1938,2 at a rate that decreased from
November until a floor was apparently reached in June. Durable
goods, especially equipment goods, for which demand temporarily
ceased altogether, of course, suffered most.3
Profits behaved similarly and, hence, until the last quarter of 1937,
conform better to expectation. As far as data of quarterly reporting
1 Weekly operating rate of steel was at about 70 per cent, or above that, from the middle
of April on, about 75 per cent from the middle of September on. It touched 80 in
December.
2 Steel and automobiles kept up well during the first 8 months of 1937, while cotton,
wool, leather, and other lines declined already during the first half of the year. Then the
iron and steel index dropped by 65 per cent and the automobile index by over 50 per cent
within 4 months. Production of electric power, which had increased strongly in 1935 and
still more so in 1936, surpassed the latter record until August and then fell precipitously.
The operating rate of the steel industry, 90 per cent in March and April 1937 (see above)
was down to 25 per cent at the beginning of January 1938. It began to rise at the begin-
ning of June.
3 Department-store sales, which from the middle of 1935 to the middle of 1936 had risen
by about 16 per cent, were well sustained, and even at the end of the year not much below
the 1936 peak; rural sales did still better and were higher for the second than for the first
half.
THE WORLD CRISIS AND AFTER 1015
concerns allow us to judge, profits were larger by one-half in the first
quarter of 1937, and lower by about one-third in the fourth, than they
had been in the corresponding quarters of 1936. For the year as a
whole they were higher by about 7 per cent.1 Of the 700 concerns
9.6 per cent reported net loss. Substantial gains on 1936 figures were
shown by steel, railroad equipment, machinery, agricultural implements,
electrical equipment, oil, metal, and mining including copper and copper
products, the automobile industry being among the chief mourners.
Factory employment rose and fell much less than output all along.
In the monthly average of 1935 it was but little over 4 per cent above the
level of 1934, then increased at a somewhat greater rate in 1936 and
reached its maximum, a little over the 1923 to 1925 average, in the
second quarter of 1937. From July to December it fell by 14 per cent.
But it was still 84 per cent of that average early in 1938 and 79 per cent
for April. Subsequent changes in most lines were small, but decreases in
the steel, machinery, motorcar, and men's clothing industries still further
reduced the index through May. Pay rolls, of course, rose — also to
about the 1923 to 1925 average in the first half of 1937— and fell, by 23
per cent from July to December, more than employment. But the rise
above and the fall below proportionality reflect not only advances or
reductions in rates but also increased or decreased employment in indus-
tries paying higher than average wages and decrease or increase of part-
time employment, here and there even overtime — but since this was
largely due to the reduction in hours, it was but another form of increase
in rates — and its elimination.2
Money wage rates gained, as we have seen, less in 1935 than they had
gained in 1934. Then they scored further gains. Average hourly rates
for skilled and semiskilled males in manufacturing (64 cents in 1934)
were about 66 cents in 1935 and about 69 cents in 1936. Unskilled males
did worse, and women suffered a small loss in reaction to the particularly
strong increase which had previously occurred in their case. In 1937 rates
were increased by another 10 per cent, reaching a level substantially
above that of 1929, while real rates then surpassed the latter more than
25 per cent.3 Labor cost per unit of product also rose. Again, it
1 Total national income was $69 billion.
2 As always, indices of retail sales and of total wage payments moved closely together.
For example, for 1929 to 1935 the chart at the bottom of the last page of the Cleveland Trust
Company's Business Bulletin for Apr. 15, 1936.
3 Real rates were, after their rise in 1933, substantially constant during 1934, 1985, and
1936. They rose by about 6 per cent in 1937, as they should in recession. There was
further increase in the first half of 1938.
According to the National Industrial Conference Board, Wages, Hours and Employ-
ment in the United States, 1914-1936, Table 2, average wage rate of factory hands was
about 62 cents in 1936 (working hours per week about 39). This compares with a rate of
1016 BUSINESS CYCLES
follows from previous argument that this must have been a major factor
in the industrial situation which it tended to make more sensitive to
depressive influences, and a major reason why unemployment, in spite
of a substantial decline in 1936 — there was even sporadic shortness of
labor — remained at a high figure. Besides forcing the pace of labor-
saving rationalization, it may in spots even have interfered with expan-
sion of output while prosperity lasted. Construction may serve as an
example.
The strong increase in contracts awarded which we observe in the
fourth quarter of 1935 was not repeated in 1936, although both total and
privately financed construction gained considerably over the year. In
1937 publicly financed construction (a little over 1 billion) declined by
over 15 per cent, but privately financed contracts for building and
engineering work (about 1.8 billions in the 37 States) showed an increase
of almost 40 per cent over 1936, to which public utilities contributed
considerably. The index advanced at the end of the year and into
January 1938. * In February there was a sharp fall in nonresidential
construction — though this was partly made up for in March — and pri-
vately financed contracts for the quarter were 30 per cent below the first
quarter of 1936. But even before the slump, the showing was again,
as in the recovery, below what we should expect considering possibilities
(see below) and monetary conditions. Building costs, which rose to the
1929 level in 1936 — they increased further by nearly 20 per cent in 1937 —
and the role of wage rates in them obviously supply the explanation.
However, it must be emphasized once more that such increase in hourly
rates as occurred from the spring of 1935 to the last quarter of 1936 is not
only quite within expectation, a rise of money wage rates being a normal
element of prosperity,2 but also that this rise cannot be thought of as a
cause of the slump. Not even the total rise from the depression low can,
except as an element of a complex pattern (see below, Sec. G, 5, c) — here,
as always, we must beware of two- variable arguments.
2. Conditions of extreme monetary ease prevailed throughout.
Prime commercial paper, for instance, was at % of 1 per cent until March
1937. Then it managed to climb to 1 per cent, but relapsed to the old
figure by June 1938. Rates on customers' loans, which on May 31, 1935,
25 cents in 1914 (working hours per week about 51.5). Monthly average of the board's
cost-of-living index (1923 = 100) was 61.3 for 1914 and 82 for 1936, i.e., it increased by
about 34 per cent.
1 This was, however, partly due to the new building code that New York City put into
force at the end of January.
2 It cannot be objected that this is speaking from a model which does not include the
possibility of abnormal unemployment in prosperity. It does; for if the labor market be
imperfect, unemployment will not prevent a rise in wage rates.
THE WORLD CRISIS AND AFTER 1017
had been 1.83, fell to 1.67, and, after an excursion to 1.71, to 1.63 by
May 1938. In spite of the heavy treasury financing, the yield of treasury
bonds declined, after a slight rise in the third quarter of 1935, to a
little over 2.2 — a development which was accompanied by great refunding
operations from the beginning of 1935 — and the yield of AAA corporate
bonds almost without interruption to 4.5 at the end of 1935 and to a
little over 3 per cent at the end of 1936. Then there was an increase
through April, but decline to new lows in May 1937, mortgage rates
keeping again a much higher level all along. This behavior of short
rates and yields obviously calls for explanation other than can be gleaned
from our model. We cannot even see a trace of the expected effect in
the slight increase that occurred late in the upswing, since that resulted
from measures of monetary management (see below). But the irregu-
larity should not be exaggerated. We shall see it in its true proportions
if we remember the behavior of English rates from 1873 to 1896.
It must also be emphasized again that the financing of 1929 still
exerted effects which should also go some way toward mitigating surprise
at the low figures of domestic corporate issues. The monthly average
of 378 millions in 1936 measures well up to the 1925 to 1929 average, but
most of those issues were for refunding. Only the second quarter of
1937 surpasses, with 140 millions per month, the modest figure of 1931
for new capital issues.1 But under the circumstances full time at machine
and machine tool shops was perfectly compatible with that. As far as
this goes, the common saying that private investment did not really
revive during the upswing must be modified. Other Loans behaved, in
fact, more nearly according to expectation and the shift in bank assets
that is characteristic of prosperities did not altogether fail to show : other
loans of reporting member banks (101 cities) started to increase from the
end of 1935, continued to do so up to the last quarter of 1937, and then
fell, while investments, partly at least in connection with this, fell from
about the middle of 1936 and began to increase again in the last quarter
of 1937 and at the beginning of 1938. Variations in deposits were too
much under the influence of monetary policy and government action in
general to be trusted to reflect the pulse of our process, except perhaps
in the second half of 1936 (see below, 4) .
All that need be said about stock prices is that they continued to
copy fairly faithfully the course of profits, to the point of disregarding the
lowering of margin requirements that came into effect on Nov. 1, 1937,
and the relaxation of the rules about margin trading that was decreased
in December,
1 Then they fell off to 75 in the third and 45 in the fourth quarter. The figure for the
first quarter of 1938 is 87 millions and for the second, 68 millions a month. But issues for
refunding bank debts are included.
1018 BUSINESS CYCLES
If wholesale prices (B.L.S. index of all commodities) may be said,
according to the guess we ventured to make in an earlier place, to have
in 1933 made up for as much of their fall as was due to the preceding
spiral, they should then have resumed their downward course. They
would, in fact, have done so but for the rise that was forced upon farm
products and foods: the index of the prices of other commodities declined
until the second quarter of 1935.1 The slow and hesitating increase
in the latter, which then began and after a setback continued until
October 1936, was exactly what we should have expected for the pros-
perity phase of a Juglar located at this one is within the Kondratieff:
we know why in this particular pattern — compare the eighties — prices
offer a strong resistance to any forces that strive to raise them, to the
internal ones which originate in the mechanism of prosperity, as well as to
any external ones, political or other.
But precisely because of this we cannot agree with those observers
who hailed the violent rise which occurred in the fourth quarter of 1936,
as the sign that prosperity — let alone "recovery" — had come at last.
From the standpoint of our process that was, on the contrary, abnormal
and calls for external explanation, which, of course, is not far to seek.
That rise of prices heralded not prosperity but** inflation."2 The gears
1 Textiles and hides and leather in 1935 reacted by a rise to their fall in 1934, the former
not quite reaching, the latter surpassing the level obtaining at the beginning of 1934.
Metals and metal products, building materials, house furnishings hardly changed at all
until the general rise in the last quarter of 1937, and other groups, though displaying more
fluctuations, behaved similarly.
We shall in the text confine discussion to the behavior of the Other Commodity Index,
because agrarian prices were too much influenced by public policy to be relevant to our
discussion. It could be shown, however — perhaps it is even obvious — that they did not
entirely fail to reflect our process. We will recall that farm products followed up their
spectacular rise in 1933 by an almost equally strong one in 1934 and the first quarter of
1935. From May 1935 to May 1936 they, on the whole, declined; but cash receipts from
sales for 1935 were, nevertheless, well above those for 1934. From June 1936 to March
1937 we have another spectacular rise in the index of farm products, resulting in a still
larger cash revenue from sales, and for 1937 the latter again increased by about 7 per cent,
in spite of the precipitous fall which, from April 1937 to March 1938, carried that index to
below the level of the last quarter of 1934. The continuation of the back-to-the-farm
movement, which was revealed by a special survey of the division of agriculture of the
census bureau (in cooperation with the bureau of agricultural economics, 1938) is, hence,
not surprising, though interpretation must differ from that for the same movement during
1930 to 1932.
2 The meaning of that term is obvious in this case. And, however we may dislike the
word, there is no other which in this connection would be equally suggestive of a self-
reinforcing sequence of increases in monetary values, receipts, and costs that will, unless
stopped from outside, go on indefinitely, each step enforcing the next and defeating the
effects of the preceding one. The role of armament demand in the rise of some prices
(copper, scrap, and so on) as well as in other features of that prosperity (operating rate of
the steel industry) must not be forgotten, however.
THE WORLD CRISIS AND AFTER 1019
of the engine composed of public spending and newly created facilities
for credit expansion began to mesh. The case illustrates, as regards the
former, the proposition that income-generating expenditure may raise
the price level in the presence of underutilized resources and, as regards
the latter, the proposition that increasing the lending power of the bank-
ing system does next to nothing in the depression which it is intended to
remedy, and very little in recovery, but takes effect when it is not intended
to do so, viz., in prosperity. It equally illustrates the mechanism that
works by two levers, anticipations on the one hand and, what is more
important, a race between prices and wages1 on the other. Of course,
only fractional use was actually made of the powers of the deposit-manu-
facturing machine, but this is no objection to that diagnosis. And it is
significant to observe how much of it went into financing of households*
expenditure. Banks, trying to find outlets for their idle cash and
responding to the incessant appeals, underlined by threats, that they
should lend more freely, sometimes went to the length of inviting applica-
tions for personal loans by newspaper advertisements, but more com-
monly financed intermediate lending agencies and retailers' receivables,
from installment paper to open accounts. Cash lending by other types of
lenders also increased. Thus, consumers* credit which had been reduced
by depression rose again to the 1930 figure or nearly so. We are not going
to reopen the theoretical problem of consumers* credit. It is enough to
point to the fact, its relation to prices, and its importance in any relapse,
which it may be sufficient, long before any great percentage of households
default, to turn into a vicious spiral.
The powers of credit creation being what they were, that process
could not only have gone on indefinitely but also at a pace beyond the
possibilities of expanding physical output. We cannot, therefore,
completely rely on the automatism of our process for full explanation of
the precipitous fall of prices in the last quarter (the rise ended with the
first) that rapidly tapered off in the first half of 1938. But expectation
from our model is for a fall during recession phases, and the recession of
a Juglar that runs its course on a Kondratieff downgrade should end up
with a price level lower than that of the neighborhood from which the
Juglar rose. If the fall that actually occurred was from a higher level
and hence steeper than can be explained by our process, the eventual
result was not so different from what it should have been if nothing else
had contributed to it. We should even have expected a decline to below
1 If such a race is not allowed to start, results may be different and more akin to those
predicated by the better type of "reflation" theories. The German and Japanese cases
are conspicuous and the English case is a less conspicuous instance of the success that may
attend a corresponding policy under appropriate conditions. For further comment see
below sub 5,
1020 BUSINESS CYCLES
the figure of the second quarter of 1935. No group save farm products,
foods, and textiles fell as much as that, and the index of commodities
other than farm products and foods never returned even to its annual
figure for 1935. There is a strong presumption that, barring monetary
management and other price-raising policies — policies, that is, which are
price raising in effect, whatever the intentions and phraseologies — price
level would continue to fall, though gently, for more than another
decade.
It may be noted that of the 784 commodities which enter into the
B.L.S. all-commodity index, the 189 items that fell most from 1926 to
1933 (when their prices ranged from 5.4 to 42.9 per cent of the 1926
figures) were at the end of 1937 on the average at about 60 per cent. Of
these all but 10 belong in the categories of extractive raw material,
agricultural products, and little-processed staples. The 190 which fell
least (and in 1933 stood at between 78.1 and 118.3 per cent of their 1926
prices) were at the end of 1937 on the average at about 100 per cent.
Highly finished articles form the bulk of this group, such as chemicals,
agricultural implements, and so on. The implications of this are in many
cases weakened, in some reversed, by taking account of changes in quality.
In others, special conditions explain the "rigidity." In no case is it possi-
ble to infer from these facts alone anything about lack of balance in the
price system or about lack of flexibility in prices per service unit.1 The
relative fall in raw-material prices foreshadows, and is a condition of, a
new equilibrium at vastly increased figures of output.
3. There is thus some justification for going on to speak of Juglar
phases. Our main reason for doing so lies, however, in the nature of the
industrial processes of the period. This becomes apparent if we ask our-
selves what we should have expected to happen. We may think, for
example, of our experience with the railroad Juglars of the nineteenth
century. They all had a family likeness and were quantitatively dom-
inated by railroad construction not only as long as this was the funda-
mentally new thing but also for a time of completing development.
Historical record does not lead us to expect that innovations of the first
magnitude — in a financial sense — will turn up in Kondratieff downgrades :
every railroad Juglar had, besides marking a step in the evolution of the
railroad system, its own contribution of novelties to make; but quan-
titatively they were of minor immediate importance — even Bessemer
steel was, for instance — as compared with the innovation that made the
railroad Kondratieff. We know what made the current one and are
thus in a position to form a definite "forecast" in order to compare it
with the actual course of things. We also know that downgrades are
1 On this subject see Professor E. S. Mason's excellent article on Price Inflexibility,
Review of Economic Statistics, May 1988,
THE WORLD CRISIS AND AFTER 1021
characterized by very numerous small and induced innovations. To
these it is impossible to do justice. But the great lines are simple enough
to list.
In doing so we do not on the whole meet with disappointment at the
first step. This is, first and foremost, the Kondratieff of electricity.
The current Juglar should have carried on the work of its predecessors
at least as much as the fifth Juglar of the second Kondratieff carried
on the railroad work of the four that went before it, at least because
investment opportunity seems even greater in this case, considering the
work to be done within the range of present technical and economic
vision.1 The production of, and the innovations in, electrotechnical
manufacturing, in fact, come fully up to expectation, so obviously that
we need not stay to prove it. Kilowatt-hours produced passed the 1929
mark in 1935. Power developments have consisted chiefly in progress
with the great public ventures: Boulder Dam, Bonneville, Grand Coulee,
Fort Peck, and Muscle Shoals, which are to increase capacity installed
by over 4 billion kilowatts, and in the smaller projects sponsored under
the Rural Electrification Act of 1936, mainly by rural cooperatives.
But the amount of construction done by privately owned public utilities
is indeed disappointing. We should have expected vigorous expansion in
power plants, substations, and transmission lines, and the pecuniary
investment corresponding to this and to the incident expansion of
equipment should have contributed decisively to the processes of pros-
perity. Whether or not the failure of the actual development (which
was, however, not negligible) to bear out those expectations is to be
recorded against the principles that yield them depends on whether or not
it is adequately accounted for by some inhibiting factor external to the
industrial organism. This will be touched upon later.
In the second place, this is the Kondratieff of the automobile. No
such development as that of the motorcar industry has ever broken off
suddenly. Therefore the current Juglar should include, or partly consist
in, another automobile wave. As everyone knows, we are again not
disappointed in this expectation. Such prosperity as there was clearly
centered in the motor industry and its satellites, such as tires and inner
tubes, plate glass, steel, by-product coke, and gasoline, and this accounts
for the characteristic inequality of the upswing as between industries
which was evident even in 1936 when improvement had become general.
Automobiles (cars and trucks; United States and Canada) recrossed the
4 million unit line in 1935 and the 5 million unit line in 1937, when produc-
1 In 1985 only 800,000 of the 6,800,000 farms were supplied with electricity. Urban
domestic consumption was 678 kilowatt-hours per home, whereas according to Mr. Samuel
Ferguson it might be 8,400 kilowatt-hours (?). It increased by 50 per cent between 1920
and 1936.
1022 BUSINESS CYCLES
tion was higher than in any previous year excepting 1929, or, if we take
account of the decline in the number produced for export and merely
consider production for domestic consumption, about 94 per cent of the
1929 figure. It is true that owing to the slump in the second half of the
year part of this output merely went into dealers' stocks, but on the other
hand, it must be remembered that part of the increase in the motorization
of the country occurred in rural communities (82 per cent of farmers
owning cars according to the Bureau of Home Economics of the Depart-
ment of Agriculture), the demand from which was to two-thirds satisfied
by second-hand cars. The essential point, however, is that the industry
was not simply, as we have expressed it, "drawn along** by environ-
mental growth or improvement (roads) and that it not merely grew into
existing, but also created new economic space. The changes in the
product were not merely routine changes in design and so on, such as
occur in operating any textile mill, but included also a number of, if
individually minor, innovations — 33,721 patents were issued in this field
since 1934 — such as the turret top (all-steel frame), automobile radios,
knee-action suspension, nonfading finishes, and others, among them still
"incubating'* ones as, for instance, the pancake motor.
Moreover, the industry or some of its concerns reached out into neigh-
boring fields and became responsible for innovations in these — the Diesel-
engine division of General Motors (new plants in La Grange and Detroit),
which had much to do with the increase in speed of trains during the
period, and the same concern's activities in the fields of refrigeration,
air conditioning, small-scale light arid power production, and aviation
(Delco Frigidaire and Conditioning Division, Delco Radio Division, Delco
Electric Light and Power Plants, Bendix Aviation Corporation, North
American Aviation) may be cited as examples. Considerable outlay for
current developments, as well as for new and improved plant capacity,
attended this development. The " competing-down " process is evi-
denced by the increasing share in production of the three leading concerns.
Behavior of stock prices, wages, and prices of product conform to our
general idea of an industry that was still innovating and expanding under
the impulse of innovations. Average hourly rate excluding increase in
payment for overtime increased, for example, by nearly 20 per cent in
1937 and was then over 28 per cent above the 1929 level.1 As to price,
there is the usual difficulty about quality.2 Retail sales value in 1937
was roughly 3.85 billions as compared with about 4.77 billions in 1929.
The allied industries all display, though to a varying degree, the
same characteristics. For the rubber industry, in particular, every one
1 In the first half of 1938 the average rate was a little over 90 cents an hour. The
working week was (April, 1938) a little over 31 hours.
8 Discounts and trading-in allowances, moreover, make quoted prices all but meaningless.
THE WORLD CRISIS AND AFTER 1028
of the above statements could be paralleled.1 Innovation was of a
similar type (" supertwist " again, the tractor and implement tire, output
of which in 1936 was 1,775 per cent above that of 1933, activities in
rubber and even in cotton growing and cotton milling), but price and
effect on output must be corrected for the increased amount of service
units contained in a modern tire of good quality — according to an esti-
mate by the industry on an average 31,446 miles in 1937 as compared
with 18,546 in 1929, which would make a fall in price per mile of some-
thing like 38 per cent. Wage rates also make a similar showing.
Improved quality and more economical use progressively deprive
steel output of its value as a cyclical thermometer.2 Only if this is taken
into account, do the 33.4 million tons of ingots produced in 1935 or the 46.9
millions produced in 19363 acquire their true significance and some com-
parability with the figures of the preceding Juglar prosperity, and only
then is it possible to recognize the output of the first five months of 1937
as truly indicative of Juglar high tide.4 The relative importance of the
lighter steel products, steel sheets, strip wire, tin plate, and so on which,
though they enter into farm implements and machinery in general, may
be said to be more nearly consumers' goods' material was, of course,
greater than at any previous time — absolute output was also at record
figures in 1936 and during the first three quarters of 1937 — but its varia-
tions yet indicate the course of cyclical subphases: lighter products were
47 per cent of the total in 1935, only 42.5 per cent in 1936 when prosperity
had got into its stride, and about 43 per cent in the first half of 1937
when it was tapering off. The new advance in rolling ("continuous
mills"), foreshadowed at the threshold of the depression, constituted
the most conspicuous innovation, but there was a large number of smaller
1 According to the Automobile Manufacturers' Association's estimate, sales value of
automobile-tire shipments (including casings, inner tubes, solids, cushions, and sundries)
was 298 millions in 1932, 307 millions in 1933, 324 in 1934, 384 in 1935, 436 in 1936, and
thus in 1936 still below not only 1929 (722) but also 1930 (532).
2 But against the facts that a pound of steel, as we have put it elsewhere, goes in general
much further than it used to even in the twenties, and that the share of high-quality special
steels and steel alloys increases, must be set the other fact that an ingot yields a smaller
amount of finished product in the case of this high-quality steel. Iron ore consumed is,
of course, influenced by the rapid increase in the use of scrap.
3 Capacity per head of population was then at the lowest figure since 1929, but for the
reason mentioned this also does not mean what it seems to. Its maximum occured in
1934, after which it was reduced by wholesale scrapping, another indication of rapid
technological advance.
4 But even apart from those considerations, ingot output of January 1937, when the
industry operated at 86 per cent of capacity, was nearly at record high, since it was only
exceeded by the output during a few months of 1929, which was clearly exceptional.
Moreover, it should be remembered that also after 1873 peaks of pig-iron production
remained below that for 1872 for quite a while.
1024 BUSINESS CYCLES
ones, principally in the field of alloys and other specialties (flat-rolled
steel), but also in others, some of which, such as the progress in welding,
were effective in creating new markets. The amounts reported as spent
by the industry on new construction and equipment, a little less than 700
millions for 1935 to 1937, do not indicate more than that there was non-
negligible investment, for that figure, on the one hand, includes some
replacements and mere extensions while, on the other hand, additional
allowance should be made for improvements financed under other head-
ings. New blast furnaces — three in 1937, two of them replacing, though
of course with improvements, dismantled ones — coke ovens for a million
tons beyond replacement, open-hearth furnaces, electric furnaces, and
then the new rolling mills mentioned constitute the main items. Behav-
ior of prices, employment, and wages also conforms to expectation.
Composite price of steel (American Iron and Steel Institute), which was
$67.71 per ton for 1923 and $47.41 for 1933, increased to $56.85 in 1937,
average hourly rates were respectively 59.6, 52.4, and (from Mar. 16,
1937) about 83 cents, well above the industrial average and above 1929.
Employment (number of wage earners; not man-hours) increased in
spite of the labor-saving nature of some of the innovations, and was about
30 per cent above the 1929 level in the summer of 1937 (over three times
the number in 1879, which is, however, a census figure and not entirely
comparable) .
In appraising steel developments and prospects, armament and
construction demand must be taken into account. * The latter comes in
as a negative item, i.e., the cyclical significance of steel production must
be interpreted in the light of the fact that this component failed to
contribute as much as we should have expected. We have noticed the
fact and one of its causes, but an additional remark suggests itself. We
would not per se have expected a particularly high wave of residential
building during a prosperity phase, although this is more likely to occur —
as it did occur in 1925 and 1926 — in the downgrade than in the upgrade
of a Kondratieff. But we miss enterprise which in this case there were
particular reasons to expect. The mass production of the cheap,
prefabricated house is one of the most obvious innovations of the present
and near future, and it should if anything have been promoted by high
1 It is, presumably, also necessary to take account of the fact that the demand for
railroad equipment, which materially contributed in 1984 and after, was greater than was
warranted by the results and prospects of the railroad business, which continued throughout
recovery and prosperity to illustrate our competing-down process. Operating revenue
from freight of Class I railroads, which at its maximum of 1929 was only 10 per cent above
1920, fell to below 60 per cent in 1932 (minimum) and was not quite 75 per cent at the end
of 1936. Revenue from transportation of passengers fell all along except in 1923, reached
80 per cent of 1920 in 1932, was still lower in 1933 and only very little over 30 per cent at
the end of 1936 (below 30 per cent for the year).
THE WORLD CRISIS AND AFTER 1025
building costs. Such enterprises first emerged at the threshold of the
current Kondratieff (City and Suburban Homes, New York, 1896,
Washington Sanitary Improvement Company, at the same time) and
prefabrication, domestic electrification, steel developments, and other
innovations1 are so many propelling factors. Yet nothing of that kind
happened on any scale, at least until the passage of the Federal
Housing Act in 1936 setting up a Federal Housing Administration
which was authorized to "insure" mortgages for projects up to 10
millions.
The chemical industry does not disappoint us. Expansion, innova-
tion, and investment were on a considerable scale — advance in the fields
of "synthetic organics," of refrigerants, of protective coatings (which
also gave an impulse to the paper industry), or of plastics may indicate
a type of improvements which sum up to a very substantial item — both
within and without the two big concerns. The rayon industry out-
stripped previous records partly by conquering new uses or markets and
partly by technological progress, which included a major innovation.
The standard fiber did not even begin to show its possibilities — and
threats. New in our sense was air conditioning. Although installations
started in 1919, they reached a maximum value of but 17 millions in
1930. The industry was still in the experimental stage. It was only in
the prosperity of the current Juglar that it rose to the modest heights of
35 millions in 1935, of 53 in 1936, and about 85 in 1937.2 Aviation may
be said to have reached about that stage in which railroads were in the
thirties of the nineteenth century. Aircraft making enjoyed a great
boom every year, 1938 also included, setting a new record, mainly,
however, because of military demand, which accounted for about 60
per cent of 1937 sales and for most of the highly profitable exports which
amounted to 34 per cent of 1937 sales. Even so, the decisive tech-
nological advance achieved hardly began to unfold its effects, and total
sales (109 millions in 1937) and profits of the eight major concerns were
distinctly modest. Still more so was the progress, in everything but
equipment and quality of service, of the three great operating companies,
whose operating revenue for 1937 did not exceed 25 millions, spelling
deficit in all three cases. Two-thirds of their total investments of about
120 millions would ex visu of that year have to be considered as lost.
Many reasons besides the peculiar difficulties incident to this industry
(and the series of disasters) account for this. But since its lack of
1One is even fundamentally new, the "cotton house." Mechanization of the actual
building was also further developed, but there was no new departure there. Improvement
in assembling parts still lags behind improvement in the parts themselves.
2 (Incomplete) estimates from figures of principal firms by LaRue Applegate, The
Annalist, Feb. 12, 1937, p. 268.
1026 BUSINESS CYCLES
quantitative significance for the current Juglar is obvious, we need not
touch upon its problems.
4. Having thus satisfied ourselves that the processes which in the
past used to carry prosperities have not been absent in the present
instance, we have established a right to speak of a Juglar prosperity and
to infer from experience that it would have asserted itself without any
external impulse being imparted to the system by government expendi-
ture or any other factor. In particular, there is nothing %to indicate that
objective opportunities were smaller or capitalist motivation weaker
than they had been, say, in 1925. The problem why that prosperity
was so weak, and why it should have been followed by so severe a slump
now emerges in its proper setting.
At the outset we dismiss the possibility — which, in fact, has not been
sponsored, so far as the writer knows, by any economist — that the steps
taken toward freer trade (Montevideo conference of American states and,
partially in fulfilment of the pledge there made, the Reciprocal Trade
Agreements Act of June 12, 1934, and the Presidential Proclamation of
July 8, 1935) can have materially dampened prosperity or intensified
depression. Of the 17 agreements actually entered into up to March
1938, only those with Cuba (September 1934) and Canada (January
1936) can possibly have exerted nonnegligible effects in spots,1 and from
what depressive effects they did exert it would be necessary to deduct
gains in other sectors. There may even have been a net contribution to
recovery and prosperity. If so, it cannot have been significant, however.
Exports, which in 1929 had amounted to 5,241 millions, rose from the
1932 figure of 1,611 only to — in millions of full-weight (or devaluated)
dollars— 1,280 (2,133) in 1934 and to 1,370 (2,283), 1,474 (2,456), and
1,977 (3,295) in the subsequent years. And the effects of general improve-
ment in the world, of devaluation, and of armament demand account
for the bulk of that.
It is less obvious that the momentous changes in the sphere of money
and credit, which have been the subject of so much controversy, did not,
except by facilitating government expenditure, decisively influence
economic processes during the years under survey. In order to establish
this, we need not go into the principles involved in those changes or into
the question of what long-run results are likely to be in the future. We
have indeed noticed the effects on money rates, on price level, and on
consumers' credit. But the lessons administered by an experience
1 Even effect "in spots" is in some cases doubtful. American shoe producers were
probably no unbiased judges of the dangers that threatened from the Bata concern (at
least, inasmuch as it remained outside the United States) or steel producers of the danger
to their lives from an "increase in imports of 173 per cent" in the first year of the Belgian
agreement, which turns to have been 1.5 millions.
THE WORLD CRISIS AND AFTER 1027
extending over a century and a half and the scarcity of acceptable
applications for credit — admitted, in the end, even by fact-finding
agencies set up for the purpose of convicting banks of restiveness — pre-
vented excesses which in fact were not at the time encouraged by the
federal reserve authorities.1 The Banking Act of Aug. 23, 1935, codified
and made permanent the chief innovations previously introduced, but
in doing so it emphasized the restrictive rather than the expansive
element in them. Following upon the removal of the restrictions
previously imposed upon transactions in foreign exchange (Nov. 12, 1934)
it was widely understood in that sense, in spite of the criticism it met from
"sound-money" quarters.2 Later on, it was proved that — at least for
once — the brakes of the engine were not mere window dressing. The
only problem relevant to our subject, therefore, concerns the use which
was made of them and which many people held responsible for the occur-
rence or the severity of the slump.
Faced with an influx of gold, as persistent as it was natural — since,
barring all extraeconomic factors, which of course helped, any commodity
will go to where it is being overpaid — and perhaps somewhat concerned
about the violent rise in prices, the Treasury and the Federal Reserve
Board took action. The Treasury, not prepared to go back upon devalu-
ation, entered at the time of the French devaluation (Sept. 28, 1936) into
the Tripartite Agreement with England and France, undisclosed oper-
ations under which helped to control the gold movement temporarily.
Moreover, in December 1936, it inaugurated the gold sterilization plan: by
directly acquiring and impounding newly received gold it prevented,
1 See, for example, the "conservative" reply of the board of governors of the Federal
Reserve System to the chairman of the Senate Committee on Agriculture and Forestry,
published in the Federal Reserve Bulletin of September 1987.
2 The act did allow the small state banks outside the system to enjoy insurance of
deposits without joining. It did nothing about the division of responsibility for monetary
policy between the Treasury and the Board. It increased rather than diminished the
Board's independence, such as it was. It developed the tool of open-market operations,
sanctioned the Board's power to vary reserve requirements, and otherwise strengthened its
hold on the policy of member banks. It failed to include the clause in the House bill about
"stabilizing business." It retained the gold cover requirements for Federal reserve notes
and deposits. It also retained, although it made them illusory, eligibility requirements.
It no doubt encouraged long-term lending by member banks, but not more than it had
been encouraged before. It perpetuated deposit insurance. Much irrelevant discussion
arose from confusions between the technically unsatisfactory working of eligibility require-
ments and the principle which they faultily expressed; between responsible practice of
member banks and laissez faire; and between their functions in the economic process and
such things as their "duties to depositors." Fundamental truth was often clothed in
wrong and fundamental error in correct arguments. Solicitude for public morals and
public welfare often vested the garb of group interests, antisocial tendencies and levity
the garb of " progress! veness." On the whole, however, the practice envisaged by the act
as it stands is still serious banking.
1028 BUSINESS CYCLES
up to September 1937, any effects of the influx on bank reserves and
deposits, the increase in monetary gold stock and treasury cash going
into an Inactive Gold Account — an important, if negative, measure.
The Board, using its new powers, directly operated upon the incubus
of excess reserves by raising reserve requirements, first by 50 per cent
(effective Aug. 16, 1936) and then again by 16% and 4^f per cent
(effective Mar. 1 and May 1, 1937). Total and excess reserves had been
rising steadily from the beginning of 1933 for all other, and from the
beginning of 1934 also for New York City member banks* thus absorbing
much of the flood and once more illustrating the value of the theory that
banks are always "loaned up." Great as was the gold influx in 1934, it
was surpassed in 1935 by more than 50 per cent (total net gold imports
roughly 1.75 billions, over half of them from France). Together with
various smaller items which need not concern us, this spelled an increase of
nearly 1.9 billions in the monetary gold stock — an increase of about
350 millions in money "in circulation" and an increase of about 1,5
billions in members* reserves, or of about 1 billion in their excess reserves.
Influx continued, though with abated force and significant intermissions,
throughout 1936: the monetary gold stock rose by more than 1 billion and
members' reserves by almost as much, the loss to circulation being almost
wholly compensated by the reduction of federal balances at reserve
banks. In 1937 the tide swelled again, 1.3 billions producing an increase
in monetary gold stock of 1.5 billions, in spite of the reversal of the move-
ment in November and December. But, as stated above, owing to the
sterilization policy this did not go into bank reserves, which continued,
however, through 1937 to move at the high level attained toward the
end of 1936, not far from 7 billions for all member banks. The increase
in deposits (adjusted demand deposits; all reporting member banks) of
about 2.5 billions in 1935 and of about 1.7 in 1936 substantially reflects
government disbursements and, until May 1936 investments, thereafter
other loans.1 They began to fall at the end of the first quarter of 1937
and went on falling almost uninterruptedly, though more in New York
than Outside, throughout that year. The first part of this development
until about the middle of 1936 does not call for additional comment; the
second part brings us back to our problem.
In our survey of time-series contours it has been observed that pros-
perity did not bring about that stiffening of money rates which, even in a
prosperity within a Kondratieff depression, we should have expected.
The facts we have just glanced at — excess reserves, in particular — amply
explain this. It has also been pointed out that, given these conditions,
the microscopic rise of some rates — commercial paper rate not among
1 There was a slight contraction, for the 100 cities outside New York more than slight,
at the beginning of 1936, which was quickly made up for and is cyclically understandable.
THE WORLD CRISIS AND AFTER 1029
them — that occurred in 1936, and the somewhat more perceptible though
still insignificant increases that occurred in 1937 also in bond yields,
cannot primarily be explained by the modest increase in business borrow-
ing but was due mainly to monetary management. It was, in fact, a
reaction to the increase in reserve requirements — under conditions of
imperfect competition curtailment of even an unsalable excess of supply
can have effect on price.1 But while monetary management produced
this effect, it certainly did not, via the rate of interest, produce any other;
for no business calculation can in practice be affected by so minute an
increase.
But together with the increase in all other loans, tne increase in reserve
requirements by 50 per cent was the signal for member banks to start
reducing their investments. The effectiveness of this signal can, how-
ever, have been due only to the fact that member banks had for other
reasons already begun to feel uneasy about their portfolios. For no
reaction was in itself enforced by that measure, which only reduced excess
reserves from about 2.9 billions (July 15) to 1.8 (Aug. 19), because almost
simultaneous treasury disbursements absorbed the effect to the extent
of about 360 millions.2 It is true that application of this weapon, as
shaped, will unavoidably create some difficulties unless the reserve posi-
tions of all members are exactly equal. But these difficulties were negli-
gible in this case, and very few members had to borrow small amounts
from reserve banks at the critical time. Even withdrawals of funds from
New York correspondents did not amount to much, nor did the sales of
investments between the middle of July and Aug. 19 (160 millions) and in
November. Adjusted demand deposits declined but slightly in August
and then increased vigorously again.
It is a question of some interest why the Board did not leave things
at that. The prospective reduction in federal income-generating
expenditure was, after all, no secret, and pro futuro the gold influx was
being taken care of by the sterilization policy. The revival of business
borrowing, which kept up total earning assets of members and even
increased outside deposits, and the strong increase in money in "circula-
tion" were no reasons for further action, considering the phase of the
industrial process. Mechanistic views about the Supply of Money seem
once more to have carried more than their due weight. But even so, the
announcement on Jan. 30 of a further increase in reserve requirements of
33^<j per cent cannot be held responsible for any depressive symptoms
and, especially, for the rapid fall of corporate security issues in the third
1 This argument, of course, applies also to the upward shift in business demand for
loans. But it is reasonable to suppose that the conspicuous measure of raising require-
ments was the more potent factor.
2 That is why excess reserves were at about 3.25 billions on Aug. 15.
1030 BUSINESS CYCLES
and fourth quarters of 1937. In case the behavior of interest rates should
not be considered sufficient proof, we will observe how well the market
stood the experiment. Some management was indeed necessary, especially
at the second step, and transition was then not quite so easy as it had
been with the 50 per cent increase. In April, banks prepared themselves
for that second step, which, as it seemed at the time of the announcement,
should have reduced excess reserves to 500 millions, but reduced them
only to about 850, owing to an increase in total reserves, the decline in
deposits, and government disbursements. The reserve system also helped
by an open-market purchase (96 millions), and all this raised excess
reserves to 1.6 billions toward the end of April. Thus prepared, the
market went through the treasury financing and the quarterly tax pay-
ments of June without any difficulty. Not even New York banks
(which one would think should have displayed some symptoms of strain)
had to borrow from the reserve bank. Bill dealers reduced rates on bank
acceptances on June 22 after 200 millions of treasury bills had on June 17
and 18 been repaid without replacement. There was some interbank
borrowing, but to a very small extent. It is true that members liquidated
investments to the amount of about 2 billions, counting from the peak in
1936 to the end of September 1937 — all other loans reached a peak in
October and then declined — and that this naturally weakened the market
of governments. But this was, as our narrative amply shows, not due to
strain, nor did it cause any. It was, moreover, what happens in every
prosperity without producing any depression.
In deference to the prevailing belief in the gratia efficiens of the supply
of money, those steps toward normalization of money-market condi-
tions were retraced precipitously. Both Board and Treasury went to
entirely unnecessary lengths in the opposite direction when the slump had
developed. While it is not correct to say that what monetary manage-
ment has proved itself able to learn was how to create slumps, it is correct
to accept that panicky retreat as an indication of the practical value of
brakes, the application of which so quickly results in excess of speed. The
Treasury first desterilized 300 millions of gold in September,1 then reduced
the sterilization policy to a shadow by the decree of Feb. 14, 1938, and
finally reversed it by releasing on Apr. 14, 1938, the whole of the 1.4
billions of gold in the inactive account, i.e., practically by transferring the
1 This might have been enough to put a stop to divestment by member banks, since
it taught them that benevolent powers would always step in to relieve them of any danger,
however remote, of illiquidity. As a matter of fact, however, the discontinuance of sales
of governments by members which then occurred, also coincides with a beginning decline
in loans, which then fell by about % billion within the year. Reversal of the downward
tendency in investment was as much contingent upon this as upon the release of gold, or
more so.
THE WORLD CRISIS AND AFTER 1031
equivalent to the Treasury account with federal reserve banks. There-
upon, the Board, which had already authorized further open-market
operations— kept within modest limits, however — and reduced margin
requirements on security transactions in November, also reduced reserve
requirements by 13.25 per cent1 on Apr. 15, 1938. Members* total
reserves having risen toward the end of 1937 because of that release of
those 300 millions of gold, the increase in monetary silver, the open-
market operations, and later also because of some purchases of gold,
excess reserves were more rapidly built up again. They were 2.5 billions
by Apr. 20, 1938; 2.75 for June; and went to 3.15 in July.
Treasury redemption of bills — which was the chief use so far made of
the formerly inactive gold — sent their yields to practically zero. Adjusted
demand deposits moved in the opposite direction to total loans plus invest-
ment, increasing by 768 millions,2 while the latter fell by 250 in the second
quarter of 1938 and cash piled up. These measures were more recently
supplemented by advance on the other line of expansionist policy, viz.,
by new rules for the lending practice of member banks, calculated to
"liberalize" it. Thus we have before us, untouched by previous experi-
ence but implemented by much more powerful tools, the main ideas of the
policies of 1933-1934. No doubt, argument based on the logic of post
hoc ergo propter hoc will be heard and read very soon. It should be
noticed, moreover, that some of the steps taken cannot be retraced, that
the rest can be retraced only with difficulty, and that, as we have put it,
such a policy, while ineffective in depression, tends to become viciously
effective afterward. Monetary policy per se may, hence, become a major
factor in the near future; but it had but little to do with the prosperity
of 1935 to 1937, and nothing with the subsequent slump.
Finally, we will return to income generation by means of government
expenditure and try to appraise the effects of its cessation in 1937. Our
historical survey (Chaps. VI and VII) has supplied us with instances of
crises that occurred near the upper turning point of a Juglar. Although
recession is not depression, the transition from prosperity to recession,
implying as it does a difficult reorientation, always creates some danger of
breakdown. We have, it is true, also seen that severe slumps at or near
that point are without exception associated, not simply with excesses of
speculation, but with excesses of speculation induced by a supernormally
rapid pace of industrial development, conditions which were conspicu-
ously absent in the case before us. But the idea suggests itself neverthe-
1 Reserves against time deposits were reduced from 6 to 5 per cent, against demand
deposits to 22.75 per cent for central city banks, to 17.5 per cent for reserve city banks,
and to 12 per cent for country banks.
2 Velocity, of course, fell rapidly and by the middle of 1938 was about 25 per cent below
its maximum annual figure (1936).
BUSINESS CYCLES
ess, that government expenditure may have played the role which in the
ast belonged to the expenditure by innovating firms and that, consider-
ng its quantitative importance, its cessation acted in much the same
manner as the cessation of the latter did in previous cases. The elements
of truth in this argument are no more obvious than its limitations.
There can be no doubt not only that income generation by government
must always create problems of adaptation, but also that in this case its
timing was singularly infelicitous.1 Its high-water mark oame exactly at
the time when the economic process could most easily have done without it
and its cessation exactly at the time when the economic process was in its
most sensitive phase. Reference to our experimental count will illustrate
this drastically. The widespread opinion that the cessation of that
income generation was the "cause" of the slump thus derives some sup-
port from our analysis,2 although it should be superfluous to add that no
argument for permanent deficits follows from that.
But the explanatory value of the shock thus imparted to the system
should not be exaggerated. Since no excessive expansion or speculation
had been called forth by the preceding spending policy — any more than
by such "natural" prosperity as there was — and since there was, as we
have seen, no monetary strain, the usual starters of downward spirals
were not operative. The injections were not suddenly discontinued but
tapered off gradually. Hence, the analogy with previous crises that
occurred around the upper turning point fails after all. The obvious
inference is that the collapse induced by the cessation of income genera-
tion was so severe, and that the jolt which it would in any case have been
natural to expect turned into a slump, because under the surface watered
by income generation the processes characteristic of recession in our sense
(see Chap. IV) failed to work as they always had worked before. The
parachute refused to unfold.
5. a. This view, which implies the presence of an additional and more
fundamental problem, is hardly controversial and, in fact, shared by
very many fellow economists who, as the reader knows, offer in explana-
1 So was the disposal of the amount spent. Sums were thrown about in the country
almost at random, without systematic regard to existing structures and probable develop-
ments, thus creating industrial and commercial positions that rested on nothing but this
temporary flood of money and were bound to die off when it ebbed.
2 Since the payments under the Social Security Act materially helped in (slightly more
than) balancing the combined federal cash account, it is even true to say, as has been
said by some economists, that those payments had for the time being a "deflationary
effect." As a long-run proposition this would, of course, not be true. Another point
should be noticed in passing. The downturn of 1937, following upon that of 1930, impinged
upon a business community which for the time being was supernormally "crisis conscious."
Moreover, many concerns may have harbored vivid recollections of what the "stand"
they had made in 1930 had cost them.
THE WORLD CRISIS AND AFTER 1083
tion the theory of vanishing investment opportunity.1 The vogue that
this theory enjoys in this country is obviously due to the occurrence of
that slump after what is universally recognized as an abnormally weak
prosperity (or "recovery"). Since in order to understand the economic
situation of our time — and much besides — it is essential to realize fully
why that explanation cannot be accepted, we will restate the case against
it at the risk of repetition.
The validity of that theory is not denied on the ground that its basic
proposition is wrong. This proposition can in terms of our analysis be
rendered as follows. Capitalism is essentially a process of (endogenous)
economic change. Without that change or, more precisely, that kind
of change which we have called evolution, capitalist society cannot exist,
because the economic functions and, with the functions, the economic
bases of its leading strata — of the strata which work the capitalist engine
— would crumble if it ceased: without innovations, no entrepreneurs;
without entrepreneurial achievement, no capitalist returns2 and no
capitalist propulsion. The atmosphere of industrial revolutions — of
"progress" — is the only one in which capitalism can survive. Hence
the capitalist organism cannot, in case opportunities for innovations give
out, settle down into a stationary stage without being vitally affected,
as it could if "changes in production functions" were an incident to its
life process and not the essence of it. In this sense stabilized capitalism
is a contradiction in terms. Moreover, it is reasonable to expect that
this kind of stabilization would produce a class of abnormalities and
instabilities of its own. There would be increasing reluctance to invest
or even reinvest, a tendency to "live on capital," to hold on to balances,
to recreate vanishing returns by all the shifts open to a class which,
though by then economically functionless, yet would, like its feudal
predecessor, for a time retain the powers acquired by and associated with
the functions previously filled. Maladjustments, unemployment and
underutilization of resources — though now of a different nature — and
neutral, unstable, and subnormal equilibria might hence well stay with a
nonexpanding world.3
1 For an admirable exposition of that theory, see A. H. Hansen, Full Recovery or
Stagnation, 1938.
2 This, of course, is putting it strongly. But the necessary qualifications must be
supposed to be by now familiar to the reader, as must also the argument summed up in that
statement. Jt should be observed that what we have called Growth is not separately
mentioned. The reasons for believing that the class of phenomena thus designated is not
important for the argument which is to follow will presently become obvious.
8 Such a world would not display cycles in our sense: a cycle in a nonexpanding world
is also a contradiction in terms. But fluctuations of the type which we have called waves
of adaptation (Chap. IV) would continue for some time. And the circumstances alluded
to in the text may, while transition lasts, also cause fluctuations of yet another type.
1034 BUSINESS CYCLES
The colors of this picture will fade if proper account be taken of the
fact that transition to a stationary state would not be sudden but would
necessarily come about by slow degrees (see below). Also, it must be
borne in mind that the proposition in question is not as a rule formulated
in that way. Some writers try to demonstrate it by means of models
which assume unchanging production functions or "methods of produc-
tion " and thus exclude the pivot on which it turns. Most writers unduly
stress the mere mechanics of the saving-investment process. But as far
as the result is concerned, there is nevertheless affinity between their
view and the one submitted in this book. We may even admit that one
of the difficulties of that transition may proceed from people wanting to
invest and getting ready to invest, while they are not able to do so at rates
of returns acceptable to them (see below). It will, therefore, be con-
venient to accept the current slogan for the purposes of the argument in
hand. The connecting link which allows us to do so is the fact that
innovation is, directly and indirectly, the great source of investment
opportunity.
Nor do we take issue with the companion proposition that investment
opportunity in this sense may, and in fact is quite likely to, vanish some-
time in the future. The reasons usually offered for this are old acquaint-
ances of ours.1 For instance, although we hold that the conquest of the
air may, entrepreneurially speaking, be as important as or more important
than the conquest of India and that, from the standpoint of our analysis,
the two are exactly the same kind of thing, we shall not deny that oppor-
tunities of the latter type are being, or ultimately will be, exhausted.
Or, although we hold that nothing at all reliable can be predicated about
it, we do not deny the possibility that technological innovations may
some day give out, either "objectively" or because people do not care to
proceed with the available ones.2 We have even an element of our own
to add. The mechanization of "progress" (Chap. Ill) may for entre-
preneurs, capitalists, and capitalist returns produce effects similar to
those which cessation of technological progress would have. Even now,
the private entrepreneur is not nearly so important a figure as he has been
in the past. We have, moreover, noticed (Chaps. VII and XIV) the
implications of chemical and other developments which may result in
making innovation capital saving or at least less capital absorbing than,
say, it has been in the railroad age. Also, it may well be true that an
1 See, in particular, Chaps. Ill, IV, and IX (retardation).
2 It should be repeated, however, (see Chap. IX) that any assertion about this would
be pure speculation. So would any assertion to the effect that the innovations which
have materialized so far were more important, more profitable, or more capital absorbing
than those that are to come. Such assertions repeat Malthus's methodological error with a
vengeance.
THE WORLD CRISIS AND AFTER 1035
increasing proportion of the "things still to be done" will lend itself to
public rather than private enterprise,1 although this would per se not
mean more than the addition of another component to a tendency
toward public enterprise which exists independently of it.
Finally, we do not even exclude the possibility that investment oppor-
tunity might vanish through saturation. The argument from declining
birth rates, in particular, loses but little by the fact that it is often inade-
quately formulated. Reduced "need" for expanding capital equipment
in a society in which population increases at a decreasing rate — still more
in a society in which it is stationary or declining — is not the point. It
does not matter whether the purposes to be served by an expansion or
reorientation of the productive apparatus strike the observer as partic-
ularly "necessary" or not — whether it is radio sets or cradles that are in
demand. And so long as the majority of people in the civilized, let alone
the uncivilized, nations are as far from anything like saturation2 as they
are now, no shrinkage of total investment opportunity will result from
the saturation of a particular want — even if we disregard the temporary,
but for the time being very important, effect of the changing age distribu-
tion. But there is a more fundamental objection which at this state of
our analysis should be familiar to the reader: "needs," whatever they
may be, are never more than conditioning factors, and in many cases
mere products of entrepreneurial action; it is not they that set the capital-
ist engine into motion, as the old household examples (China and so on)
show; and economic development (capital consumption included) has
never been conspicuous in the countries which to the observer seem to be
most lavishly supplied with needs. However, the argument may be
1 The criteria according to which it is to be decided whether a given proposition is
economically more fitted for public than for private enterprise vary, of course, historically
and are at present blurred by the prevailing preference for the former, which is largely
extra-economic. But we need not go into this beyond reminding ourselves that there
is a class of cases in which inadequacy of private profits indicates inadequacy of social
advantage.
2 Once more we meet that overemphasis on the possibilities of expansion within existing
lines which we repeatedly met before. But it should be observed that even these are
important enough to negative any idea of deadlock from that source for quite a time.
Another argument may be noticed here. It has been held that the process of providing
"capitalist" equipment for the population — the stock of producers' goods necessary for
the Boehm-Bawerkian "roundaboutness" — was a historically unique and uniquely capital-
absorbing task, which has been performed once for all in the nineteenth century. This
seems not only to assume that, within the production functions that existed at any time,
capital equipment has been carried to saturation point, but also to overlook that the inser-
tion of new production functions in many cases, not to say typically, annihilates the old
equipment economically, so that that task has to be periodically repeated and, as experience
abundantly shows, repeated by means of new savings and new credit creation. Long-
distance transportation by motorcar, "replacing" long-distance transportation by railroads,
is surely not financed from the depreciation accounts of the latter.
1036 BUSINESS CYCLES
upheld, at least to some extent, in another way: as has been pointed out,
provision for an indefinite family future is of central importance in the
scheme of bourgeois motivation, and much driving power may be elimi-
nated by childlessness.
b. But we do take issue with the third proposition, which asserts the
relevance of those considerations for the diagnosis of the situation of
1938. Obviously, we have been speaking of longest-run tendencies just
now. Opportunities for technological or organizational qr commercial
innovation cannot be thought of as vanishing (if they are vanishing)
except very gradually. If there actually be a general tendency toward
decline in capital absorption, it can assert itself only in time, though
shocks to individual industries may be both sudden and serious. The
rate of increase in population declines imperceptibly per year. The
call of entrepreneurial adventure is too deep-seated to cease dramatically
of itself. And so on. Such tendencies, even if well established as some
of them undoubtedly are, qualify but ill for the task of explaining the
peculiarities of a particular Juglar. They may affect contour lines over
time and bend them downward. But they cannot explain the weakness,
relatively to its predecessor, of any given prosperity, and they look
absurd in the role of explaining factors of a sudden slump. If it be held,
nevertheless, that one or all of them suddenly acquired dominating
importance at any given historical juncture, such an assertion requires,
in order to be taken seriously, proof not only of a secular tendency or
"trend" but of the presence of circumstances adequate to account for so
improbable an occurrence as sudden action would be.
It has been observed above that the essentially gradual modus operandi
of those tendencies must enter into any speculations about the phenomena
to be expected from vanishing investment opportunities. For instance,
there is no warrant for the assumption which has been made the basis of
far-reaching conclusions, that in the face of them people will go on saving
at a rate sufficient to produce difficulties. Owing to the persistence of
habits, this could conceivably be so in the short run of a depression phase,
though we have seen plenty of reasons to doubt it. But in the short run
investment opportunity cannot decline perceptibly. And in the long
run there is no reason to suppose that savings — and both things and
psyches in general — will not adapt themselves. Moreover, it is obvious
not only that declining rates of savings will accompany declining birth
rates because both phenomena flow from the same sociopsychological
source, but also that there is a causal connection between the two.
But reasons less general than that estop us from accepting the theory
in question. Whatever may be thought about those "trends" and the
way in which they operate, none of them has in this country advanced
far enough to bear it out. We have seen in some detail that "objective "
THE WORLD CRISIS AND AFTER 1037
opportunities are not lacking. We are less than ten years removed from
as vigorous a prosperity as was ever witnessed and from a depression
provably due, in the main, to the pace of preceding "progress." It has
been argued that that prosperity differed in character from previous
ones by the prominence of (durable) consumers' goods production, and
hence indicated that a fundamental change had, already then, occurred
in the cyclical process. Of course that prosperity differed from the Jug-
lar prosperities of the Kondratieff upgrade. But it did not differ in
character from the comparable Juglar prosperities of the preceding
Kondratieff downgrades, and therefore does not indicate any fundamental
change in the working of the capitalist organism. Expansion of pro-
duction of consumers' goods, including expansion in the fields of utilities
and of public works, was, proportions guarded, no less prominent a
feature in the developments of the twenties, thirties, seventies, and
eighties of the nineteenth century. Nor can it be urged that funda-
mentally new opportunities of first-rate magnitude are not in prospect.
Barring the question whether that is so, it is sufficient to reply that in
the eighteen-twenties hardly anybody can have foreseen the impending
railroad revolution or, in the eighteen-seventies, electrical developments
and the motorcar. No less an authority than John St. Mill1 compromised
himself by holding in 1870 that the possibilities of capitalist enterprise
were substantially exhausted. Vestigia terreant.
As applied to the American situation of today and to the abnormalities
of the current Juglar, the theory that the capitalist process is stagnating
from internal causes inherent to its logic and that income generation by
government is nothing but the self-defense of a shriveling organism, is
therefore a complete misfit — at best a mistaken interpretation of certain
aftereffects of the world depression, at worst the product of wishful think-
ing on the part of all those who crave for a presentable basis for policies
they approve. It still retains two merits, however. The one con-
sists in the many elements of truth which, as we have seen, enter into its
arguments as distinguished from its application. The other consists in
the recognition, by implication at least, of the fact that, as any social
system depends for its functioning and survival on the actual delivery of
the premia it holds out, so capitalism depends for its functioning and sur-
vival on the actual delivery of the returns, anticipation of which provides
its motive power.2 For this is, after all, what the stressing of investment
1 This has been pointed out to the writer by Mr. R. Abel-Musgrave.
* See Chap. XIV, Sec. C, 2. This would in the times of intact capitalism have been
taken for granted. In our time, however, the attempt to run capitalism in an anticapitalis-
tic way has given rise to arguments which came near to denying it. Hence the recognition
by the theory of vanishing investment opportunity of that rather obvious point may well
be recorded to its credit.
1038 BUSINESS CYCLES
opportunity amounts to. Slightly extending and modifying the mean-
ing of that phrase, we may hence continue to use it ourselves and
agree that it is vanishing investment opportunity which is the matter
with present-day capitalism — anything can, in fact, be put into that
form, the structural principles of the capitalist process being what
they are. And our task then reduces to substituting for unconvincing
reasons why investment opportunity should be vanishing, a more
convincing one.
c. The analysis of Chap. XIV, Sec. B supplies it: capitalism produces
by its mere working a social atmosphere — a moral code, if the reader pre-
fer— that is hostile to it, and this atmosphere, in turn, produces policies
which do not allow it to function. There is no equilibrating apparatus to
guarantee that this atmosphere or these policies should develop in such
a way as to prevail in the fullness of time, i.e., when the capitalist process
will have really spent its force or be spending it. Whenever they prevail
sooner, there is danger of a deadlock, by which we mean a situation in
which neither capitalism nor its possible alternatives are workable. This
is what, to a certain extent and presumably not yet for good, has happened
in this country.
It might be replied that anticapitalist attitudes are also, like the
tendencies adduced by the theory of vanishing investment opportunity
in its usual acceptance, a matter of slow growth and, hence, similarly
open to one of the objections raised against that theory above. But we
are able to do in this case what cannot be done for those tendencies, viz.,
to show that, and how, that attitude came suddenly to a head and
suddenly acquired dominating importance. And anticapitalist policies,
unlike attitudes, may be dated.1 The coincidence in time between
them and disappointing performance of the economic engine is indeed
striking. We will survey them under the headings of Fiscal, Labor, and
— for want of a better expression — Industrial policies.
At least since 1932 the burden of direct taxation imposed upon that
part of the national revenue which goes to the higher and highest brackets
1 It should be observed at once that, anticapitalist measures being, of necessity, meas-
ures hostile to private investment opportunity, the accredited exponents of the theory of
vanishing investment opportunity must perforce agree with the argument that is to follow;
for the consequences of inadequate investment opportunity are obviously independent of
its causes. They will be the same whether these causes are internal or external, i.e.,
whether the process itself, by virtue of the law of its life, produces inadequate margins,
or these margins, if produced, are or would be taken from recipients, or, finally, the anticipa-
tion of them is in other ways prevented from having its normal effect. Hence, those
economists will in any case, even if unconvinced by our argument against their explanation
of vanishing investment opportunity, at least have to insert ours into their schema. If
they do insert it in a place appropriate to the importance of its constituent elements, there
will not be much room left for difference of opinion.
THE WORLD CRISIS AND AFTER 1039
was undoubtedly high enough1 to affect "subjective" investment
opportunity or, as we have previously expressed it, to shift the watershed
between "to do and not to do." No other than direct or mechanical
effects need, however, be attributed to this burden until roughly 1933-
1934, because the increase in taxation was then accepted as a sacrifice to
be made in a national emergency, as it had been during the war. But
from the revenue act of 1934 on, this was no longer so. Permanence of
the burden for reasons unconnected with emergency, involving a transfer
or redistribution of wealth which in the highest brackets amounted to the
socialization of the bulk of private income, and in some cases taxation for
taxation's sake and regardless of insignificance of results for the Treasury,2
then became part of an established policy, the general drift of which was
not reversed in 1938. Some outlines of the theory of the subject have
been presented in the preceding chapter. Aspects other than effects on
the process of economic evolution are irrelevant to our purpose. The
quantitative importance of the change to the interests concerned is
unquestionable and unquestioned. Hence, we need not go into details or
follow up the successive steps embodied in the revenue acts from 1934 to
1937, but can confine ourselves to the following comments.
As the above suggests, the writer is inclined to stress the importance of
the income, corporation, and estate taxes at the expense of others which,
being novelties, have been more widely discussed. The facts that the
limit of exemption from the income tax is very high, the flat rate very low,
and the surtax distinctly moderate up to an income of $30,000, are irre-
levant to the argument. It is above that range, principally within a
group of not more than 30,000 or 40,000 taxpayers that, the structure of
American industry being what it is, those taxes, raised within a few years
to their present figures, exert a serious influence on "capital supply"3 and
1 For a careful evaluation of that burden, including an estate tax equivalent and
covering the years 1924 and 1927 to 1933, see M. Yaple Sweezy, Direct Taxes by Income
Classes, American Economic Review for December 1936.
2 The insignificance of financial results is very striking in the case of the estimates —
which are what is relevant as regards purpose — of additional revenue from the revenue
act of 1935. The increases of the surtax were, for example, to yield 45 millions more, the
graduated income tax 37 millions, the excess profits tax 10 millions, the increase of the
estate tax plus the gift tax 101 millions — the nibbling of a mouse at the mountain of the
deficit. The arguments that this was a matter of that budget and that in future booms
much higher yields could be expected are not to the point. The latter, moreover, begs
the fundamental question.
3 The writer does not wish to stress, under the circumstances of the past years, short-
run effects on quantity of monetary capital and its rate of increase. As far as these go,
that fiscal policy may even have had a net result favorable to prosperity and unfavorable to
depression through enforcing an increase in total expenditure. That element is primarily
stressed in their interesting study on Economic Consequences of Recent American Tax
Policy, Supplement I to Social Research, by Professors Colm and Lehmann, who attempt
1040 BUSINESS CYCLES
business behavior, which, of course, is greatly intensified by the failure of
legislation to permit the carrying forward of business losses,1 by the new
treatment of personal holding companies and by other inroads into actual
or potential capital.
The so-called capital gains tax has been held responsible for having
accentuated, if not caused, the slump. The writer is, however, unable to
see that it can have had much to do with the processes of the current
Juglar except by way of damping speculative ardors and thereby making
issues of stock more difficult than they would have been. %The financing
of the positive phase cannot, considering the abundance of cheap money,
have been seriously interfered with by this; the subsequent slump should,
if anything, have been mitigated by it. Other points, in particular the
effect it exerts — not by reducing " over-saving" but — by enforcing dis-
saving, though relevant to a prognosis of the results to be expected in the
future from the capitalist engine and not substantially affected by the
modifications introduced by the revenue act of 1938, need not concern us
here.
The antisaving theories and the ressentiments of the day found a very
characteristic expression in the special surtax on undistributed corporate
income (undivided profits tax), which ranged from 7 to 27 per cent.
Discarding again the question of the long-run effects which the measure
might have had if it had been allowed to remain on the statute book, we
may split immediate effects into those of a further increase in the burden
on corporate income and those of the specific penalty imposed on corporate
accumulation, and confine ourselves to the latter. It possibly resulted in
an absolute and relative increase in distributed income which is neither
certain nor easy to evaluate because there were also other reasons for the
increase that actually occurred, but which, whether great or small,
presumably increased, or helped to counteract the decrease of, system
expenditure. Nevertheless, the measure may well have had a paralyzing
influence on enterprise and investment in general. The actual presence of
accumulated "reserves," and the possibility of accumulating them
quickly, strengthens the position of a concern with respect to the risks and
chances of innovation and expansion which it confronts. One of the causes
of the efficiency of private business is that, unlike the politician or public
quantitative evaluation of the difference made to the supply of capital. But to the writer
supply in the sense of "willingness to sell," that is to say in this case, willingness to invest,
seems to have been the more immediately important thing affected.
1 That feature, for which no rational argument has ever been offered to the writer's
knowledge, is more important than it seems. A loss which can be carried forward without
penalty is one thing, a loss which cannot, quite another thing. A risk which it may be
rational to take in the first case will frequently have to be refused in the latter. This would
not, of course, apply to "small" or even moderate taxes.
THE WORLD CRISIS AND AFTER 1041
officer, it has to pay for its mistakes. But the consequences of having to
do so are very different according to whether it risks owned or borrowed
" funds," or whether a loss will only reduce surpluses or directly impinge
upon original capital. Adequate book reserves are as necessary a requi-
site as adequate stocks of raw material, and in their absence, or with
reduced facilities of acquiring or replenishing them, an entirely different
and much more cautious business policy would impose itself. In pros-
perity, investment opportunities would be seen in a perspective of
reduced proportions; in depression, firms would have to bow more readily
to the storm. In the latter case in particular, the important class of
considerations — pure business considerations among them — that used to
induce many firms "to make a stand " for some time, even at considerable
immediate loss, would tend to vanish from the businessman's mind. All
this, it is true, vanishes from the economist's mind as soon as he buries
himself in the mechanics of aggregative theory. But many industries,
which are among the chief economic assets of the nation and of which the
automobile industry is the standard instance, would under a regime of
undivided profits taxes never have developed as they did. And as regards
the current Juglar, the actual course of events both during prosperity
and during the slump is compatible with the opinion that this attack on
the foundation of corporate finance weakened the former and intensified
the latter to a non-negligible extent.1
It should be observed that this is a matter of value of assets and not of
liquidity, which under the conditions prevailing in this country since
1931 was never a problem for a concern of unimpaired standing. Simi-
larly, the argument that accumulations make it easier for a concern to
live in depression and to "cushion" the effects of depression on the eco-
nomic process by keeping up dividend and wage payments cannot be met
by pointing out that only a part of total accumulations is held in cash or
near-cash and that the rest cannot be "paid out." It is true that from
the standpoint of the individual management liquidity constitutes an
advantage. The ease with which the bulk of American large-scale indus-
try steered through the vicissitudes of 1931 and 1932 was to a consider-
able extent due to it. It is also true that accumulations which are held in
a liquid form tend to work in an anticyclical sense. But this must not be
1 An exponent of antisaving views disposes of that opinion as "ballyhoo.** After
recording his belief that the country does not want additional investment and that, if it
did, funds could easily be secured by borrowing, he asks: "Why did it (the undivided sur-
plus tax) not produce this sad effect (the slump) sooner?" No serious economist has, to
the writer's knowledge, held that the slump was solely caused by this tax. But if timing
of effects carries any weight (the reader knows that we attach but limited importance to
time sequences), it is nothing short of ideal in this case. The act went into force in June
1986. Its effects were due to show themselves urbi et orbi in the first quarter of 1987.
And at the end of the second the first symptoms of impending difficulties appeared.
1042 BUSINESS CYCLES
confused with the point which the writer is trying to make and which, in
this case, is entirely independent of cash considerations, though it would
not be so for other times and countries.
The effect on the combined federal cash account of the method chosen
for financing the social security program has been mentioned before. No
further attention need, hence, be paid to the money-market and expendi-
ture aspects of the payments into the Old Age Reserve Account and the
Unemployment Trust Fund. Independently of this, the tax on pay rolls
levied on firms was of course a nonnegligible element in the'increase of the
total fiscal burden which occurred in 1937.1 The question of effects raises
difficult problems in transference. In a situation in which wage rates are
firmly upheld and prices of the products of "big business" not allowed
to rise, increase of the tax to the full amount ultimately contemplated
may not only produce additional unemployment, but also be sufficient,
as comparison with corporate net earnings shows, to cause paralysis in
some industries, such as would, for example, enforce the creation of
another and much more stringent NRA. But for the time being no major
effect can be attributed to this tax taken by itself.
Labor policies reduced investment opportunity — besides employment
per unit of output — mainly by forcing up wage rates. We have seen,
however, that not all of the increase which actually occurred can be attrib-
uted to those policies; and precisely because rising rates were to a
considerable extent met by labor-saving rationalization, the effect on
investment opportunity was presumably not very great.2 Costs incident
to employing labor were also increased in other ways. And more than
elsewhere it is here necessary to define investment opportunity widely and
to take account of the less measurable effects of increasing difficulties in
operating plants which the growth of a new body of administrative law
entails. A major measure of this kind, the National Labor Relations Act
(July 5, 1937) was placed on the statute book in the period under survey.
As the reader knows, more vigorous use was immediately made of the
1 The steel industry, for instance (American Iron and Steel Institute; identical concerns),
paid for 1937 156.6 millions in federal, state, and local taxes, or about 40 per cent of net
earnings or 15 per cent of pay rolls, or about 60 per cent more than in the preceding year.
About 37 of the 60 per cent are accounted for by the increase in the pay roll tax.
2 It will be recalled that as far as meeting rising wages in that way involves additional
outlay on plant and equipment, investment opportunity may temporarily be increased.
Such facts as that in the steel industry the composite price of products was in July 1938 a
little below the 1923 to 1929 average, while average hourly money earnings of labor were
higher by 33 per cent must be interpreted in this light. It, nevertheless, remains true
that the course of wages was one of the factors that account for that industry's failure to
repair its damaged financial structure, as evidenced by the fact (Iron and Steel Institute)
that at the end of 1937 its surplus and reserves were still but little more than half of what
they had been in 1920.
THE WORLD CRISIS AND AFTER 1043
facilities created by it than is suggested by its actual contents, which
keep within the most ordinary lines of labor policy in modern democracies
and only develop the principles of earlier legislation, such as the labor
clauses of railroad acts, the Clayton Act, certain acts passed during the
war, Sec. 7a of the recovery act, and other enactments. Official support
given to the campaign of the Committee for Industrial Organization and
lending to the act a color not naturally its own, must be listed inde-
pendently. But after the fullest allowance for these and other elements
of the case, we shall still be left with the result that labor policies — more
precisely, what has actually been done in the field of labor policy — were
not, taken by themselves, of decisive importance in shaping the business
situations of those years.
As regards what we have called industrial policies unfavorable to
investment opportunity — or, what economically amounts to the same
thing, to action being taken on any "objective" investment opportunity,
whether declining or not, that may have been present — two instances will
sufficiently illustrate what we mean. First, we have seen reasons to
expect that developments in the field of public utilities would be a leading
feature of the current, as they had been in the preceding Juglar. We
have also seen that, barring federal enterprise, that expectation was not
fulfilled. The writer does not see how it could possibly be denied that in
this case existing investment opportunity was prevented from having its
normal effect, not so much by what was actually done, but by the blanket
threat behind it. Expected competition from federal or municipal power
plants was a factor in some sectors.1 The Public Utilities Holding Com-
pany Act endangered the American solution of the fundamental problem
of power finance. But the decisive element of the situation was that
indefinite threat: executives and investors would have had to be com-
pletely blind to the political forces that were being marshaled against
them, if they had been prepared to take the responsibility for, or to cooper-
ate in, new investment on a large scale. The case thus serves to show not
only how unrealistic any theory of investment opportunity is which leaves
the political factor out of account, but also how easily the latter may
acquire an importance compared with which that of any decline of invest-
ment opportunity from reasons inherent to the capitalist process would be
negligible, even if it did occur at a significant rate per year.
1 Even if it had been still more important than it was, this factor would have to be
listed with a qualification. The development of new sources of power and their competition
with the old ones is, of course, part of our process and not an impediment to it. On princi-
ple, it is immaterial by whom the new sources are developed, whether by public or private
enterprise. The right of that factor to a place in the list of spokes in the economic wheel
rests with the extent to which the federal and municipal works can be expected to be
privileged and to be made tools of attack irrespectively of economic (cost-accounting)
rationality.
1044 BUSINESS CYCLES
Second, there is nothing surprising in the fact that under the circum-
stances the no less old hostility against "monopoly power'* should have
asserted itself again all over the industrial field. But "monopoly " really
means any large-scale business. And since economic "progress" in this
country is largely the result of work done within a number of concerns
at no time much greater than 300 or 400, any serious threat to the
functioning of these will spread paralysis in the economic organism to a
much greater degree than a similar threat to the corresponding number of
concerns would in any other country. No compensation was afforded by
the federal government's extreme anxiety not to show hostility to private
business in general or to do anything that could have aroused the cry of
Government in Business, because the contributions of the favored strata to
"progress" and their investments are not only comparatively small but
also, to a large extent, induced by what happens in the world of big busi-
ness. That hostility propelled or facilitated the fiscal and labor policies
which we have glanced at above. Beyond these very little was actually
done; but much was foreshadowed at various times, even before the
monopoly investigation, recently instituted. This may have meant noth-
ing or everything, according to whether or not the threats — no doubt,
again indefinite — were taken seriously by those whose decisions they
could have influenced. But it should be observed how very much like
"liquidity preference owing to vanishing investment opportunity" the
behavior would look which would result if they were.
d. It will be seen that none of all the measures mentioned under
our three headings can, if considered individually, be reasonably held to
have played a dominant role in shaping the current Juglar. An easy
road thus seems to lead toward the conclusion adopted as a matter of fact
by many, if not most, economists, that no explanation can on these
lines be derived for the lack of vitality displayed by the economic process
during the period under survey, and that investment opportunities must,
hence, be vanishing from causes internal to that process, in spite of all
we have adduced to the contrary. It is, however, suggested that the
following considerations greatly strengthen the case for the adequacy of
that explanation.
First, the combined effect of a series of measures unfavorable to invest-
ment opportunity can evidently not be evaluated by adding up the effects
which each of them would have had in the absence of the others. We have
repeatedly met cases of this kind. For instance, the reader will recall that
our discussion of the course of German wage rates did not result in any-
thing amounting to proof — or rather that it led us to deny — that this
element would have spelled serious disturbance if it had acted alone; but
that, since it did not, this was not the relevant question to put. Simi-
THE WORLD CRISIS AND AFTER 1045
larly, we might in the case before us make even larger concessions than the
writer would be prepared to justify, to the prevalent tendency to under-
estimate the effects of any or all the individual measures we have glanced
at, and nevertheless have to conclude that their combined effects were
adequate to produce the observed result. The individual measures
obviously tended to reinforce each other. "Objectively0 — i.e., irrespec-
tively of intentions harbored by any individuals1 — they amounted to
systematic attack on investment opportunity all round: it was frontally
attacked by direct reduction of revenues — or the operative part of total
net revenues — through taxation, which would have been only the more
effective if there really had been also an inherent tendency for investment
opportunity to shrink; simultaneously, it was attacked in the rear by
increasing costs; and both attacks were supplemented by a third — the
attack on those traditional methods of management, pricing, and financ-
ing in the sphere of "big business" which were associated with the latter's
emergence and successes. No doubt, opinions still may in all fairness
differ as to the importance both of these combined attacks and of the
precise points in the industrial structure that were being attacked. But
difference of opinion is not possible about the relevance of the principle
of interpretation which the writer is trying to stress.2
Second, the mistake involved in trying to arrive at an estimate of
combined effects by that process of addition is not more serious than the
mistake of confining attention, in evaluating either isolated or combined
effects, to the wording of enactments, congressional declarations of policy,
and statements of the chief executive. Economists who pride themselves
on the practical bent of their researches could really be expected to know
that the personnel and methods by which and the spirit in which a measure
or set of measures is administrated, are much more important than any-
thing contained in any enactment. We have met with examples above.
The events surrounding the National Labor Relations Act will again
serve to illustrate that simple truth, particularly if we compare American
with English experience in that field : it should be obvious that in the one
1 It is hoped that the reader will not object to the above statement on the ground that
he is quite sure that Mr. X or Mr. Y does not harbor any hostile intentions. If, however,
this objection were raised, the writer would not reply by asking how the reader knows that
or by pointing out that policies are hardly ever made by the men at or near the top of the
official ladder; he would reply that intentions are entirely irrelevant. History is a record of
"effects" the vast majority of which nobody intended to produce.
2 It is neglected to a surprising degree, even in cases in which there is no aversion to
the conclusion implied. Authorities on accounting and taxation habitually discuss
effects of fiscal, authorities on labor discuss labor policies in isolation, even if their intention
is to prove the presence of "injurious" effects, apparently without realizing that they
thereby reduce their arguments to exercises in general theory.
1046 BUSINESS CYCLES
case effects on investment opportunity may result which it would be
absurd to expect in the other. This already covers part of what we desig-
nate by the term Social Atmosphere.
But, third, this atmosphere should also be listed independently as
an additional factor in its own right. It is surely not too much to ask
economists to realize that behavior in human societies differs from behav-
ior in animal societies or in physical systems, in that it not simply reacts
to "disturbances" but to interpretative and anticipative — correct or
false — diagnoses of them. Real or supposed drifts and trends may count
as much as or more than facts, threats as much as actions, indefinite
threats more than specific ones, in creating the psychic environment in
which the nation's work has to be done. We know that behind those
measures, administrative acts, and anticipations there is something much
more fundamental, viz., an attitude hostile to the industrial bourgeoisie
which is no ephemeral composite of individual circumstances and political
exigencies of the day but the product of the same social process that
created that bourgeoisie. Businessmen presumably do not hold that
theory. But they need not hold any in order to realize that there is in
those measures and programs more than there would have been in similar
measures and programs 30 years ago. They are not only, but they feel
threatened. They realize that they are on trial before judges who have
the verdict in their pocket beforehand, that an increasing part of public
opinion is impervious to their point of view, and that any particular
indictment will, if successfully met, at once be replaced by another.
Again, we may differ in our estimates of the importance both of this factor
and of the functions it tends to paralyze,1 but it should not be overlooked.
Fourth, the effects of all that on investment opportunity — if the reader
prefer, on what to the businessman appears as an investment opportunity
of a given degree of attractiveness — were greatly enhanced by the sud-
denness of the change of scene. In order to convince ourselves of this, it
is only necessary to reflect that any major change in the relations between
1 Those are, of course, different things. Concerning the first, the writer believes that
observations are available which point to the conclusion that the influence of "atmosphere"
on the behavior of entrepreneurs and "capitalists" was considerable but admits that the
nature of the case precludes proof and that any attempt at proof by consequences would be
circular. Concerning the latter, few people will doubt the importance of those effects if
the argument is made to turn on aversion to investing. But difference of opinion would at
once appear if it were suggested that effects of the hostile atmosphere can also consist in
impairing the efficiency of entrepreneurs and managers. For, as pointed out in Chap. XIV,
Sec. C, 2, it is part of the creed of the day that those people who have any reason to feel
threatened do not fill any function anyhow. That this is an article of creed and not based
on any familiarity with the facts of business practice any teacher of economics can easily
establish for himself: the belief is present at a stage in life in which nobody can have
acquired an idea of what it means to run, let alone create, a large-scale concern.
THE WORLD CRISIS AND AFTER 1047
the individual and the state, including any major shift in favor of the
latter of the shares in total private revenue earned, involves changes in the
fundamental habits of mind, the attitudes to life, and the valuations at
least of those who are immediately concerned. The sociology of this
need not detain us. But as a matter of history it is clear that such changes
usually come about by small installments and as the result of a slow
process of education, which must be far advanced for codification of
principles into a new body of law to be a success. We observe, in fact,
that the modern principles of English taxation took about 30 years to
develop and "sink in," and that the beginning of the modern system of
English social policies dates back to at least the eighties of the nineteenth
century, when the ideas of Chamberlain and Dilke spread dismay among
their colleagues in Gladstone's second administration.1 The English
bourgeoisie was thus given time to acclimatize. This is why, as the reader
will recollect, we have not in the English case used the factors under dis-
cussion in any explanation of the features of any individual Juglar phase,
or of any short-run phenomena in general, but only in our explanation of
long-run contours and levels.
But in this country there was no such preparation; hence, there was a
different reaction.2 Barring the war intermezzo, there was nothing
except the feeling against "monopolies" and utilities to indicate any
resentment, and that was of the middle-class type only and easy to keep
in hand. On the whole, the businessman's moral world was the nation's
moral world right up to the crisis. And for nearly two years the demo-
cratic administration, though doing many things which were felt to be
"unorthodox" by its friends as well as by its foes — measures actually
were, as we have seen, not so very much out of keeping with American
tradition — in no way displayed the attitude that we are discussing now
but, on the contrary, signs of a thoroughly bourgeois attitude. The
change in policy dates only from 1934—1935. It therefore followed rather
than preceded the radicalization of the public mind, which in conse-
quence of the crisis had occurred between 1930 and 1933 as radicalization
in countries in which authority is associated with military values will
occur in consequence of military defeat.
The analogy with the German breakdown in 1918 suggested by the last
remark indicates the line on which we should explain how and why a
1 It would, however, be more just to date from the preceding Disraeli administration
and, in another sense, from the forties (Lord Ashley).
2 Recognition of this element is often implied in current talk about "lack of confidence,"
"strike of capital," and a "government-made depression." These and similar phrases are
objectionable on several grounds and indicative of naive resentments rather than of
correct diagnosis, but in the impression they are intended to convey there is an element of
truth.
1048 BUSINESS CYCLES
secular process, after having failed to assert itself to any practically signif-
icant extent for fully forty years after the closing of the frontier, then
suddenly became the dominant factor of the political situation. In
doing so, we should no doubt have to go into many circumstances peculiar
to the American environment in general and to American politics in
particular, in order to understand the details of the change in attitude and
of the resulting political pattern. But the fact, the broad causes, and the
effects on business behavior are sufficiently obvious to establish our point
without any analysis of details.1 There are, however, two aspects which
cannot be passed by.
On the one hand, we have insisted above on the importance of person-
nel and of methods of administration. New measures as well as new
attitudes must be implemented by a skilled civil service. In any case
they set a difficult task to even the most experienced bureaucracy. As
a rule, however, reforming governments enjoy at least the advantage of
having that indispensable tool ready at hand — in most historical instances
it grew up along with the tendencies which they represent. This
happened, for example, in England, while in Germany the regime of 1918
was able to take over from its predecessor both an excellent civil service
and a state-broken public. In this country .a new bureaucracy had
suddenly to be created. However good part of the material on which it
was necessary to draw, and however creditable, considering the circum-
stances, the performance of a great many individuals and groups may
have been, there was no experience, no esprit de corps, no clear idea even of
what civil service is and what it can and cannot do. No less inexperi-
enced— to the point of not seeing the fundamental administrative
problems at all — were the men in whose hands that unwieldly apparatus
was put. The tact, the reserve, the savoir-faire which are second nature
to a seasoned bureaucracy were alike absent. Enthusiastic individuals
and groups developed their own policies and tried to push them with
Congress and the public, scornfully refusing counsels of self-denial and
patience. In consequence, that sense of indefinite threat was immeasur-
ably increased. English policies may be felt to be equally or more
oppressive, but they are never aggressive: spectacular manifestations of
1Some of the economic causes of the sudden change of scene have, however, been
pointed out previously, among them the agrarian situation, the last epidemic among
banks, the stock exchange crash, unemployment. The main factors which are and must
remain absent from our sketch are the structure and technique of American politics and
the role played by the "intellectuals." A complete analysis would also have to take
account of random configurations, of which one has been mentioned, viz., the fact that a
presidential election occurred a few months after the lower turning point, exactly at
the time at which the sociopsychological "hangover" of the depression should be at its
THE WORLD CRISIS AND AFTER 1049
aggressiveness proceed only from quarters that are — by all parties1 —
firmly held in check, and never from members of the public service. The
methods of the latter may be likened to deerstalking and tend to minimize
trouble and disturbance caused by any given measure. Administrative
methods in this country tend to maximize them and are more like those
of the fox hunt — and this makes a lot of difference.
On the other hand, sudden change, unless of the Russian type, is of
necessity imperfect change. It impinges upon a set of economic and polit-
ical conditions which are very unequally ripe. This puts advocates and
opponents of new departures in false positions, adulterates arguments,
and makes it impossible to face issues squarely. In England the question
of the employment of nonunionized labor, for instance, having been
allowed to mature, is now one of secondary importance. In this country
it cannot even be frankly put, yet it is at the bottom of much strategy
and struggle which, precisely because the issue is not ripe for decision,
must be expected to remain for some time a source of difficulties and losses
to all parties concerned. But the standard instance is the policy followed
with respect to public utilities. Here, if anywhere, there was an all but
united public opinion, united at least in its hostility to the private interests
involved. Moreover, European experience suggestively pointed toward
nationalization of power production and transmission, which could have
been carried without any shock to "business confidence " if the interests of
investors had been fully safeguarded, and with but a sharp and short one
if they had been sacrificed — always provided that there were no clenched
fists or indeterminate threats of other nationalizations to follow. Yet
it was not even attempted.2 The clenched fists and indeterminate threats
were all the more in evidence, however, and the result was, as we have seen,
to paralyze one force without substituting another. This will always
result from raising issues before they can be effectively dealt with and
illustrates what above has been described as deadlock. To deny
that this impairs the efficiency of the economic engine or, if we retain
the slogan, reduces investment opportunity, would seem to the writer
unreasonable.
1 The official leaders of the Labor party, hence, frequently find themselves described
as what in plain English would have to be called dunces and traitors. But it does not mat-
ter, and everyone knows that action upon the views of radical wings would speedily end in
discomfiture.
2 A move toward encouraging municipalities to acquire the works and distributive
systems that serve them has recently been made. It only illustrates our point, both by its
weakness and by its timing. Another illustration is afforded by the practice of the PWA
of facilitating the construction of distribution systems duplicating the privately owned
ones in the sector of the TVA by charging less than cost and by lending the remainder at
low rates to municipalities.
1050 BUSINESS CYCLES
If the above considerations are given their proper weight, there should
not be left much doubt as to the adequacy of the factors external to our
process1 to account both for the disappointing features in the current
Juglar and for the weakness of the response of the system to government
expenditure, in particular for the failure of the latter to affect investment
and employment more than it did. It cannot be proved in the sense in
which a mathematical theorem can, that the balloon shriveled, not from
causes inherent to its structure, but because the air was being sucked out
of it. It is, however, highly plausible and, after all, only what, if we clear
our minds of cant, we should expect to occur in transitional stages.
Prognosis would, in this country more than in any other, have to take
account of the likelihood that there will be intermissions or even reversals ;
of the effects of "acclimatization"; and of the fact that, if our schema is
to be trusted, recovery and prosperity phases should be more, and
recession and depression phases less strongly marked during the next
three decades than they have been in the last two. But the sociological
drift cannot be expected to change.
1 Let us add, however, for the last time, "in the narrow sense adopted for the purposes
of this book." In a wider sense those factors and the mentality or moral code behind them
are not external to the process of economic evolution but part of it, a part as essential and
unavoidable as any other and, in particular, as any "objective" shrinkage of investment
opportunity could be. The above argument would, hence, be completely misunderstood
if it were taken to imply that, "only" politics or humors being the matter, pristine vigor of
the capitalist process could easily be restored at the next swing of the electoral pendulum.
As far as that goes, the practical implications of our diagnosis do not differ much from those
of the theory of vanishing investment opportunity in its usual acceptance. Even govern-
ment spending as a permanent policy could be rationally defended on our diagnosis: the
pattern resulting from the action of inhibiting factors would in all respects be similar to the
pattern envisaged by the saving-investment theory; it would display the same lack of
resilience and the same tendency toward subnormal quasi-equilibria; in particular, it would
always produce or reproduce extensive unemployment. Therefore, government spending
would, given the general will to conserve those inhibiting factors, always suggest itself
as a remedy for shortrun difficulties each application of which would impose, under penalty
of breakdown, the application of the next dose. Fear of such breakdowns may in the end
become the dominant motive even among those who on principle are most strongly opposed
to spending policies.
APPENDIX
Description of the Statistical Material Embodied in the Charts
CHART I. ARITHMETIC SCALE
Three sine curves with periods respectively of 684, 114, and 88 months (or 57, 9j£, and
3;Hi years) and with amplitudes roughly proportional to the periods, i.e., in the relation of
18, 3, and 1, have been plotted separately. The fourth curve presents their sum.
<.>-»-©)•
(3) T = si
fA\ 1Q • /360A°_LQ ' ^860A°_L • /360\°
(4) y = 18 sm (—tj + 3 sm {—t) + sm (— t)
CHART II. ARITHMETIC SCALE
The graph represents the derivative (time rate of change) of the sum of three sine
curves with periods of 684, 114, and 88 months.
CHART III. ARITHMETIC SCALE
(1) Average Railroad Revenue Freight Loaded. United States. Weekly figures.
(2) Four-item Moving Average of Three-item Moving Average of Two-item Moving
Average of Series (1).
(3) Superseasonal Normal. This is a freehand curve drawn through the inflection
points of the moving average curves.
(4) Deviations of Series (2) from Series (8) multiplied by the "Inflation Factor."
For a discussion of the Frisch method in general the reader is referred to Chap V. For
a discussion of the Inflation Factor cf. Horst Mendershausen, Annual Survey of Statistical
Technique: Methods of Eliminating Changing Seasonal Fluctuations, Eeonometrica, vol. V,
no. 3, July 1937, p. 244.
Let
/ = Inflation Factor
Then
_ sin 2X sin 3X sin 4X
2 sin X 8 sin X 4 sin X
where
»-..?
and
p = twice the time distance between normal points
and
D = 1, the time unit in which p is measured.
1051
1052 BUSINESS CYCLES
CHART IV. RATE OF PERCENTAGE CHANGE OF PRICES. ARITHMETIC
SCALE
The rate of percentage change for each series is computed by applying to the logarithms
of the indices the formula
Dk - ^(iA* -f iAfc+0 - K2(sAHi
for annual indices or the formula
Dk « 2(iA* + iA*+i) ~ MGAjfe+i
%
for quarterly indices. The antilogs of the D items minus 100 % give the rates of percentage
change.
The rate-of -change formulae are applied to logs rather than absolute figures for reasons
analogous to those leading to the use of logarithmic rather than arithmetic scales on charts.
The importance of a change in the series is, hence, made to depend on the relative (per-
centage or geometric) movement.
(1) United States. Wholesale Prices. 1842-1913. The data have been taken by
permission from Prices by G. F. Warren and F. A. Pearson, published by John Wiley &
Sons, Inc., New York, pp. 12-13 (see also the discussion in the text and in the note on p. 14),
and built into quarterly 12-month averages of the monthly items.
(2) Germany. Wholesale Prices. 1879-1913. For the years from 1879 to 1902 it
has seemed best to use the monthly index compiled by Otto Schmitz, Die Bewegung der
Waarenpreise in Deutschland von 1851 bis 1902, Berlin, 1903. This index is an unweighted
arithmetic average of the prices, published since 1879 by the Imperial Statistical Office, of
29 articles, raw materials almost entirely. From 1898 to 1913 the author has, by permis-
sion, used the monthly index (average 1898-1913 = 100) of 10 commodities (cotton yarn,
cotton cloth, raw silk, wool (2), calf skins, pig iron (2), lead, tin) constructed by E. W. Axe
and H. M. Flinn, An Index of Business Conditions for Germany, 1898-1914, Review of
Economic Statistics, vol. VII, 1925, pp. 282-285. Quarterly 12-month averages of the two
indices were spliced together, bases being shifted so as to make them coincide as nearly as
possible from 1898 to 1902. At the time this seemed the best thing to do. In his later
work the writer availed himself of the new and incomparably more satisfactory index of
the Institut fiir Konjunkturforschung, leaving however the chart as it was.
(3) United Kingdom. Wholesale Prices. 1850-1913. From 1850 to 1870 the annual
Sauerbeck index (Prices of Commodities and the Precious Metals, Journal of the Royal
Statistical Society, vol. XLIX, September 1886, p. 648) was used with the permission of the
Royal Statistical Society. This index is a simple average (1867-1877 = 100) of 56 whole-
sale prices, almost entirely of raw materials: 22 foods, 8 minerals, 11 textiles, and 15 other
materials. Cotton and other articles are counted more than once. From 1870 to 1885
recourse has been had to the annual Board of Trade Index (see Cmd. 6955, 1913, p. 308)
which is a weighted aggregate (1900 = 100) of the prices of foods and raw materials, the
articles being weighted according to the estimated consumption during the years 1881 to
1890. From 1885 to 1913 quarterly 12-month averages were computed from the monthly
index of the editor of the Statist (cf. Wholesale Prices of Commodities in 1921, Journal of
the Royal Statistical Society, vol. LXXXV, March 1922, p. 275), a continuation of Sauer-
beck's index. The three indices were spliced together graphically on the basis of over-
lapping years.
CHART V. BRITISH PREWAR PULSE. LOG SCALE
(1) Interest Rate. 1824-1913. From 1824 to 1850 use was made by permission of
the series of quarterly average discount rates in London (best bills) as published by Norman
APPENDIX 1053
J. Silberling in British Prices and Business Cycles, Review of Economic Statistics, Supple-
ment, vol. V, 1923, p. 257 (see footnote 2, p. 242). For the most part these are the rates
which were charged by the firm of Overend, Gurney & Co. The data were built into four-
quarter moving totals. From 1850 to 1857 quarterly 12-month moving averages of the
monthly average rates charged on best bills by Overend, Gurney & Co. have been computed
from the figures in Parliamentary Papers, 1857, vol. X, Part I, p. 464. From 1858 to 1861
the quarterly average market rates on high-class bills were taken from R. H. I. Palgrave,
Bank Rate and the Money Market, p. 83, by permission of the publisher, John Murray.
From 1862 to 1913 the curve has been based on the figures of the Lombard Street rate, at
the beginning of each month, on two or three months' bills as published in the London
Economist's Annual Commercial History and Review. The four series were spliced by
means of overlapping periods.
(2) Production. 1785-1913. This is Dr. W. Hoffmann's annual index of total
industrial production, cf. Ein Index der industriellen Produktion ftlr Grossbritannien seit
dem 18. Jahrhundert, Weltwirischaftliches Archiv., vol. XL, 1934 II, pp. 396-398, the use
of which has been kindly permitted by the author and the editor, Professor Predohl. In
1785, 10 commodities are represented. See that article for commodities and weights.
Most of the 50 series finally included begin about 1800. It is believed that from that year
on approximately two-thirds of English industry is covered. For purposes of weighting,
the index is broken up into six shorter periods and also into broad industrial groups.
Indices are separately computed for producers' goods and consumers' goods. Foods,
textiles, materials, chemicals, iron and steel and machine industries, mining, transporta-
tion, and electricity are the main divisions covered. The reader's attention is invited to
Dr. Hoffmann's recent book on the subject, Wachstum und Wachstumsformen der
Englischen Industriewirtschaft.
(3) Prices at Wholesale. 1785-1913. From 1785 to 1850 Mr. Norman J. Sterling's
quarterly index (Review of Economic Statisticst Supplement, quoted above) has been used
by permission. It is a simple geometric mean of the prices of 35 commodities, mostly raw
materials (1790 = 100). Four-quarter moving totals were plotted. From 1850 on, the
series are as described under Chart IV (3). All splicing was done on the basis of over-
lapping years.
(4) Deposits plus Circulation. 1880-1913: deposits in joint stock and private banks
in Great Britain (exclusive of the Bank of England). The circulation is total note circula-
tion in Great Britain at the beginning and middle of each year. Both series were taken
from the Economist's Reports of Joint Stock Banks of the United Kingdom.
CHART VI. UNITED STATES PREWAR PULSE. LOG SCALE
(1) Deposits plus Circulation. 1870-1913. The data are call-date figures, five each
year, for "individual deposits" and "notes outstanding" of all national banks outside
New York City, taken by permission of the Review of Economic Statistics from A. A. Young,
An Analysis of Bank Statistics for the United States, Harvard University Press, Cam-
bridge, Mass., 1928, pp. 8-13. A five-item moving total centered on the third item has
been plotted. "Individual deposits" are individual deposits as reported by the Comptrol-
ler of the Currency minus clearing-house exchanges. They do not include government or
interbank deposits. Before Apr. 26, 1900, however, some banks counted deposits due to
savings banks as individual deposits. "Notes outstanding," of course, do not, except very
imperfectly, indicate variations in national bank notes in circulation.
(2) Interest Rate. Commercial paper rate. 1831-1913. The monthly data were
taken, with the permission of Colonel Leonard P. Ayres, from his chart Business Activity
and Four Price Series, Cleveland Trust Company Bulletin for July 1932. The sources
quoted by Colonel Ayres are: from 1831 to 1860, E. B. Bigelow, The Tariff Question, Boston,
1054 BUSINESS CYCLES
1862; from 1861 to 1865, J. G. Martin, Seventy-three Years' History of the Boston Stock
Market, Boston, 1871; and from 1866 to 1913, W. L. Crum, Cycles of Rates in Commercial
Paper, Review of Economic Statistics, vol. V, January 1928, p. 28. Quarterly 12-month
totals of the monthly items were plotted.
(8) Output of Manufacturing and Mining. 1868-1918. The curve is plotted on
annual figures taken by permission from Professor W. M. Persons, Forecasting Business
Cycles, published by John Wiley & Sons, Inc., New York. This index is a combination of
two indices given separately; both are weighted arithmetic averages, with 1909-1913 = 100
as base, manufacturing being given a weight of % in the combination. See W. M. Persons,
op. cit., pp. 172-173, for a description of the index and its two components.
(4) Wholesale Prices. 1831-1913. Monthly. See Chart IV (1). Quarterly 12-
month averages were plotted.
(5) Crop Production. 1870-1913. Professor W. M. Persons' annual x index, a
weighted arithmetic average (1909-1913 = 100). For description see op. cit., pp. 171-
172. Acknowledgements are again due to John Wiley & Sons, Inc., New York.
CHART VII. GERMAN PREWAR PULSE. LOG SCALE
(1) Wholesale Prices. 1879-1913. Monthly. See Chart IV (2). Quarterly 12-
month averages.
(2) Industrial Production (Manufacturing and Mining). 1860-1913. Annual.
The index of the Institut fur Konjunkturforschung, see Rolf Wagenfiihr, Die Industrie-
wirtschaft, Vierteljahrshefte zur Konjunkturforschung, Sonderheft 31, 1933, p. 58. The
index is computed from 57 series weighted according to the number of employees and
installed horsepower in 1907. The author's thanks are due to the President of the Institut,
Professor Wagemann, for permission to use it.
(8) Berlin Market Discount Rate. 1868-1913. From 1868 to 1898 the monthly data
used are those published by the London Economist's Annual Commercial History and
Review. From 1899 to 1913 monthly figures were taken from E. W. Axe and H. M. Flinn,
see quotation under Chart IV (2). They are averages of the high and low for the first
week of each month. Quarterly 12-month averages were computed.
(4) Note Circulation and Credit Accounts of Joint Stock Banks. 1884-1913. Annual.
Kreditoren plus Notennmlauf for the end of each year from Der Deutsche Oekonomist, vol.
XXIV, 1906, pp. 464 and 466; and vol. XXXII, 1914, pp. 538 and 580.
CHART VIII. NINE- YEAR MOVING AVERAGES OF PRICES. LOG SCALE
The following series were put in the form of annual averages from which were computed
9-year moving averages centered on the fifth year.
(1) United Kingdom. Wholesale Prices. 1779-1913. For description of material
see above under Chart V (3).
(2) Germany. Wholesale Prices. 1851-1913. From 1851 to 1902 the index is that
of Otto Schmitz, see Chart IV (2), which until 1879, when the index of the Imperial Sta-
tistical Office begins, is an unweighted average of the Hamburg prices of 24 raw materials.
For 1903-1912 figures were taken from the British Board of Trade Enquiry into Working-
class Rents and Retail Prices, etc., 1912, Cmd. 6955, 1913, p. 356, a continuation of
Schmitz's index.
(3) United States. Wholesale Prices. 1797-1913. This is the series described in
Chart IV (1).
CHART IX. ARITHMETIC SCALE
The original data (dots) are annual wholesale prices in the United States, 1790-1920,
from the U. 8. Department of Agriculture Bulletin 999, p. 2, the 5-year average, 1909-1914,
being equal to 100.
APPENDIX 1055
(1) A 3-year on a 2-year moving average (graphically determined).
(2) Curve through inflection points of 10-year cycle.
(8) Curve through inflection points of 22-year cycle.
(4) Curve through inflection points of 50-year cycle.
CHART X. PERCENTAGE DEVIATIONS OF PRICES FROM 9-YEAR
MOVING AVERAGE. ARITHMETIC SCALE
The annual averages of the series used for Chart VIII were divided by the corresponding
items of their 9-year moving averages and expressed as percentages minus 100.
CHART XI. UNITED STATES PRICES. 1840-1913. ANNUAL. LOG SCALE
(1) Consumers' Goods. Wholesale. From 1840 to 1890 the index (1890 = 100) was
computed by M. J. Fields from the data contained in the Aldrich report of 1893 (Committee
of Finance, United States Senate, Fifty-second Congress, Second Session, Report 1394,
Part I, pp. 91-94). It is a simple arithmetic average of the three groups: (1) food, (2)
clothing, and (3) house furnishings. From 1890 the index of the U. S. Bureau of Labor
Statistics Bulletin 284, pp. 48-49, has been used. See also Bulletin 149. The two indices
were spliced together.
(2) Foods. Wholesale. The curve is based on figures taken by permission from G. F.
Warren and F. A. Pearson, Prices, published by John Wiley & Sons, Inc., New York,
pp. 25-27. 1910-1914 = 100.
(3) Household Furnishings. Wholesale. Same source as (2).
(4) Textile Products. Wholesale. Same source as (2).
(5) Producers' Goods. Wholesale. From 1840 to 1890 the index (1860 = 100) was
computed by M. J. Fields from data in the Aldrich Report (see Series (1)). It is a simple
arithmetic average of two groups, metals and implements (excluding pocket knives) and
lumber and building materials. From 1890 to 1913 the index is the one published by the
Bureau of Labor Statistics Bulletin 284 [see Series (1)]. The two indices have been spliced.
(6) Metals and Metal Products. Wholesale. Same source as (2).
(7) Basic Commodities. Wholesale. Same source as (2). This index includes the
prices of 30 basic commodities, farm products, minerals, textiles, and the like.
CHART XII. PIG IRON CONSUMPTION. ANNUAL. LOG SCALE
(1) United Kingdom. 1854-1913. The original sources are the Annual Reports on
Mines and Quarries, issued by the Home Office and the Annual Statements of the Trade
of the United Kingdom. The series was taken from Parliamentary Papers, Cmd. 2145,
pp. 24-25, and the Seventeenth Abstract of Labour Statistics, p. 44. It is also given by
A. C. Pigou, Industrial Fluctuations, 2d ed., London, 1929, p. 386.
(2) United States. 1855-1913. From 1855 to 1870 this series was constructed by
E. M. Hoover from production figures of the American Iron and Steel Institute and imports
for calendar years. These import figures were obtained by interpolation between fiscal-
year import totals as given in the 1871 Annual Report of the Institute. Exports and
changes in stocks were ignored; but as late as 1870-1871 the total United States exports of
pig iron were only 3,480 gross tons or 0.02 per cent of the 1870 production. From 1871 to
1913 the series of estimates of the American Iron and Steel Institute, published in its Annual
Statistical Reports and based on data for production, imports, exports and (since 1874) net
change in stocks held was used with its permission. Data for 1871 to 1873, being given in
net tons, were converted to gross tons.
(3) Germany. 1860-1913. The series (also given by S. Kuznets, Secular Movements
in Production and Prices, pp. 469-470) is from the Statistisches Jahrbuch ftir das Deutsche
1056 BUSINESS CYCLES
Reich. The figures have been converted from metric tons (2,205 Ib.) to gross tons (2,240
CHART XIII. RATE OF PERCENTAGE CHANGE OF PIG IRON
CONSUMPTION. ANNUAL. ARITHMETIC SCALE:
The rates of percentage change of the three series plotted on Chart XII were calculated
by the use of the formula explained in the description of Chart IV.
CHART XIV. UNITED STATES. LOG SCALE
(1) Cotton Production. 1840-1911. The figures (for the "cotton year") were taken
from Circular 32, U. S. Department of Agriculture, The Cotton Crop of the United States,
1790-1911. Linters are included from 1899 on.
(2) Railway Freight Ton-miles. 1852-1913. Annual. From 1852 to 1887 the series
was taken by permission from Mr. Carl Snyder's Business Cycles and Business Measure-
ments, p. 238 (see also note, p. 39), published by The Macmillan Company, New York,
1927. This series gives net ton-miles of freight carried on Class I railroads estimated from
the figures of the principal lines as reported in Poor's Manual. From 1888 to 1913 the
series was taken from Railway Statistics of the United States of America, 1916, p. 99,
prepared by S. Thompson, Bureau of Railway News and Statistics, Chicago, 1917. Switch-
ing and terminal companies are included to 1908. These data refer to Class I, Class II,
and Class III railroads and fiscal years.
(3) Coal Production (Anthracite and Bituminous). 1839-1913. Annual. This index
(1926-1930 = 100) was taken by permission from G. F. Warren and F. A. Pearson, The
Physical Volume of Production in the United States, Cornell University Agricultural Experi-
ment Station Memoir 144, Ithaca, 1932.
(4) Cotton Consumption. 1862-1913. Annual. The figures are from the Statistical
Abstract and refer to "cotton years." (See Census Bureau Bulletin 166, p. 57).
(5) Building Permits Index. 1874-1913. Annual. This series (also used by A. F.
Burns, Production Trends in the United States since 1870, pp. 302-303) has been compiled
by Mr. Carl Snyder and published (from 1882 on) in his book on Business Cycles and
Business Measurements, p. 275. The writer's acknowledgments of permission to use it
are due to Mr. Snyder and The MacMillan Company. It represents an index of building
permits in from one to seven cities (seven since 1895) divided by an index of changes in
costs of construction. The item for 1879 (not given by Mr. Snyder) was interpolated by
referring to the behavior of pig iron production which closely paralleled the deflated
building-permits index at that time. The assumption was that the ratio of 1872f 879 *°
187^{ 880 was the same in the two series.
(6) Interest Rate. See Chart VI (2).
CHART XV. INDUSTRIAL PRODUCTION. ANNUAL. LOG SCALE
(1) United States. See Chart VI (3).
(2) Germany. See Chart VII (2).
(S) Great Britain. See Chart V (2).
CHART XVI. RATE OF PERCENTAGE CHANGE OF INDUSTRIAL
PRODUCTION. ANNUAL. ARITHMETIC SCALE
See Chart XV for the series used and Chart IV for the formula.
CHART XVII. PRODUCTION (AND CONSUMPTION). ANNUAL. LOG SCALE
(1) United States. Producers' Goods. 1870-1913. This index of the production of
capital equipment has been constructed by E. M. Hoover. 1890-1900 = 100. It is a
APPENDIX 1057
weighted arithmetic average of the outputs of iron ore (14 per cent), zinc (2 per cent), lead
(4 per cent), copper (13 per cent), natural gas (5 per cent), cement (2 per cent), rails
(3 per cent), steel ingots and castings (31 per cent), pig iron (12 per cent), coke (5 per cent),
ships (6 per cent), locomotives (3 per cent). The weights were based approximately on
value added by manufacture or on value of product (average of 1890 and 1900).
(2) Germany. Producers' Goods. See above, Chart VII (2).
(8) Great Britain. Producers' Goods. 1860-1913. See above Chart V (2). For
1785 this index covers only coal, copper ore, copper, tin. Iron and steel products, machin-
ery, etc. are included from 1787. But in the period for which it has been charted, coverage
extends to nine mining, two iron, steel and machine industries, six nonferrous metal and
metal goods, three chemical, one electricity, and one rubber series.
(4) United States. Consumers' Goods. 1867-1913. This index of consumption of
nondurable consumers' goods has been compiled by E. M. Hoover. 1867-1914 = 100.
Weights are roughly proportional to the values consumed (average of 1880, 1890, 1900, and
1910). The commodities included are sugar (16 per cent), coffee (7 per cent), tea(l per
cent), wine (1 per cent), malt liquors (17 per cent), spirits (9 per cent), tobacco (18 per cent),
and wheat flour (31 per cent).
(5) Great Britain. Consumers' Goods. 1860-1913. See above, Chart V, (2). For
1785, the index covered cotton yarn, woolen cloth, linen, malt, and paper; for 1787, also,
woolen yarn, silk, and sugar. In the period for which it has been charted, coverage extends
to six textile, eight food, drink, and tobacco, two leather, two paper, one lumber, and two
chemical series.
(6) Germany. Consumers' Goods. See above, Chart VII, (2).
(7) Series (I) divided by Series (4).
(8) Series (8) divided by Series (5).
(9) Series (2) divided by Series (6).
Series (7), (8), and (9) are plotted so that the average ratio for the whole period is
represented by the "normal," or 100 per cent, line.
CHART XVIII. UNITED STATES. ANNUAL. LOG SCALE
(1) Nondurable Consumers' Goods. See Chart XVII (4).
(2) Crop Production. See Chart VI (5).
(3) Production of Capital Equipment. See Chart XVII (1).
(4) Basic Production (excluding "products of the soil"). 1870-1913. This is Mr.
Carl Snyder's index taken by permission from G. F. Warren and F. A. Pearson, The
Physical Volume of Production in the United States, Cornell University Agricultural
Experiment Station Memoir 144, Ithaca, 1932, pp. 63-64. It includes coal, pig iron, copper,
zinc, tin, lead, steel, silver, petroleum, and nickel.
CHART XIX. GERMAN PREWAR PULSE IN RATES OF PERCENTAGE
CHANGE. ARITHMETIC SCALE
For a description of the series see Chart VII. For the method of computing the rates
of percentage change see Chart IV. The rate of percentage change of the discount rate is
charted on a scale one-tenth as large as that used for the other curves.
CHART XX. INTEREST RATES AND UNEMPLOYMENT PERCENTAGE.
LOG SCALE
(1) United States. See Chart VI (2).
(2) Great Britain. See Chart V (1).
(3) Germany. See Chart VII (3).
1058 BUSINESS CYCLES
(4) Great Britain. Unemployment Percentage. 1851-1913. The series was taker
by permission from A. C. Pigou, Industrial Fluctuations, published by Macmillan & Com-
pany, Ltd., London, 1st ed., 1927, pp. 353-354. The figures represent the trade unions
unemployment percentage, as published in British and Foreign Trade and Industry
(Second Series), Cmd. 2337, pp. 89-92, and the Seventeenth Abstract of Labour Statistics
p. 2. Persons on strike, or locked out, sick, or superannuated are excluded. The percent-
ages for some of the earlier years were partly computed from expenditure on unemployment
benefit.
CHART XXI. UNITED STATES INDIVIDUAL PRICES* DIVIDED
BY THE GENERAL PRICE LEVEL. ANNUAL. LOG SCALE
(1) Wheat. 1866-1913. Average farm price per bushel, taken from the Department
of Agriculture, Yearbook of Agriculture, 1914, p. 522.
(2) Rubber, Para. 1856-1913. From 1856 to 1889 the July items from the Aldricl
Report [see Chart XI (1)] Part II, pp. 291-292. From 1890 to 1913, figures of the Bureau
of Labor Statistics, Bulletin 149, p. 175.
(3) Petroleum, Crude, Barreled. 1862-1913. From 1862 to 1892 the July items from
the Aldrich Report, Part IV, pp. 1835-1836. These are monthly averages from 1862 tc
1873 except for the year 1867. From 1890 to 1913 figures of the Bureau of Labor Statistics,
Bulletin 390, pp. 134-135: annual average (Pennsylvania) price relatives on the base
1913 = 100. The items up to 1890 were shifted to the same price relative basis.
(4) Bituminous Coal. 1857-1913. From 1857 to 1891 the July items from the Aldricl
Report, Part II, p. 178. From 1890 to 1913 annual average price relatives (1913 = 100, oi
"Georges Creek," f.o.b. New York harbor for 1890 to 1912 and "Pocahontas" mine run,
f.o.b. Norfolk, Virginia, for 1913) from the Bureau of Labor Statistics, Bulletin 390, pp,
130-131. Items before 1890 were shifted to the same price relative basis.
(5) Bessemer Steel Rails. 1867-1913. From 1867 to 1889, as published in the Aldrict
Report, Part II, p. 215. From 1890 to 1913, figures of the Bureau of Labor Statistics,
Bulletin 149, p. 149.
(6) Anthracite Coal, Stove. 1840-1913. From 1840 to 1891 July items from the
Aldrich Report, Part II, p. 177. From 1890 to 1913 figures of the Bureau of Labor Statis-
tics, Bulletin 390, pp. 126-127 ("New York, Tidewater").
(7) Copper, Ingot. 1840-1913. From 1840 to 1889 July items from the Aldrict
Report, Part II, p. 185. From 1890 to 1913 annual average price relatives, 1913 = IOC
from the Bureau of Labor Statistics, Bulletin 390, p. 150 (up to 1907, "lake," and after,
"electrolytic" copper). Earlier items were shifted to the same price relative basis.
(8) Railroad Freight Receipts. 1852-1913. Average freight receipts per ton-mile
From 1852 to 1892 an average of data for a varying number of railroads given in the Aldrict
Report, Part I, pp. 615-617. From 1889 to 1913 average for years ending June 30, pub-
lished by the Interstate Commerce Commission, Annual Report on the Statistics of Railwayt
of the United States. The two series were spliced together. The 1908-1912 figures are not
strictly comparable with the rest, because they include returns from switching and terminal
companies.
(9) Pig Iron, No. 1 Anthracite Foundry. 1844-1913. From 1844 to 1890 the data
were taken by permission from J. M. Swank, History of the Manufacture of Iron in All
Ages, Philadelphia, 1892, p. 514. From 1890 to 1913 the figures of the Bureau of Laboi
Statistics, Bulletin 390, pp. 138-139, have been used.
Each of the above series was divided by the wholesale price index, see Chart IV (1),
Gold or paper price indices were used according to the nature of the individual price
quotations.
APPENDIX 1059
CHART XXII. BRITISH SHIPPING. LOG SCALE
(1) Total Tonnage in Existence in the United Kingdom, equal to steam net tons plus
one-third of sail net tons from the Statistical Abstracts of the United Kingdom. The
increase in 1914 must be discounted because the provisions of the Merchant Shipping Act of
1907 became fully operative on Jan. 1 of that year. From 1923 on, the Irish Free State is
excluded. Data are plotted as of ends of years.
(2) Annual Increase in Total Tonnage. First differences of series (1) plotted in the
middle of each year.
(3) Shipbuilding in the United Kingdom. Net steam tonnage plus one-third of sail
tonnage. Plotted as of midyear. Warships built for foreign governments are included
from 1870 to 1888, but from 1886 on there is another series which excludes them. The two
series were spliced. Both are from Statistical Abstracts for the United Kingdom. In 1897
the engineers' strike put a stop to all construction.
(4) Interest Rate. See Chart V (1).
(5) Price of "New, Ready, Cargo Steamer," 7,500 tons. Quarterly. This series was
taken by permission from Fairplay's Weekly Shipping Journal, vol. CXXXIV, Jan. 10,
1935, p. 102.
(6) Freight Rate Index. Annual. This index was made up by splicing together
indices from four sources on the basis of overlapping years.
Board of Trade figures were used from 1884 to 1903.
C. K. Hobson in his Export of Capital (London, 1914), p. 182, continued the Board of
Trade Index from 1904 to 1912.
An index from 1884 to 1924 was taken from F. C. James, Cyclical Fluctuations in
Shipping and Shipbuilding Industries (University of Pennsylvania thesis, 1927), p. 78.
This index was compiled by Dr. Isserliss, Statistician to the Chamber of Shipping, London,
and was used with his permission and that of F. Cyril James. It is based on E. A. V.
Angier's figures of ocean freights published annually in the statistical number of Fairplay's
Weekly Shipping Journal and in his Fifty Years of Freights, 1869-1919, "Fairplay,"
London, 1920. The figures represent a weighted average of freight rates to and from
England for the whole of each year, and do not give a complete picture of the years during
which freight rates changed. The engineers' strike in 1897 affected rates and so did, of
course, the Spanish American and South African wars. In 1915 rates were fixed by
government, and in 1917 all tonnage was requisitioned.
From 1920 to 1934 an index on the base 1920 = 100 was taken from The Statist, London,
Oct. 29, 1921, and later numbers. This index has also been compiled by Dr. Isserliss. It
is a geometric mean of tramp quotations for eight routes to and from the United Kingdom.
The more important routes are represented by more than one quotation.
CHART XXIII. UNITED STATES. LOG SCALE
(1) Deposits minus Investments outside New York City. 1890-1914. These are
actual call date figures for all national banks outside New York City. The deposits are
"net" deposits, i.e., individual deposits (see Chart VI (1)) minus clearinghouse exchanges,
plus amounts due to minus amounts due from other banks. The "due to" (or from)
includes amounts due to reserve agents, other national banks, state banks, trust companies,
and savings banks. These are not the "net" deposits of the Comptroller which are com-
puted for purposes of obtaining reserve ratios. The figures are from A. A. Young, op. cit.,
pp. 8-13 [Chart VI (1)]. The investments exclude United States securities held to cover
circulation and United States securities and other bonds held to cover United States
deposits. They are also actual call date figures for all national banks outside New York
City and have been taken from the same source.
1060 BUSINESS CYCLES
(2) Clearings outside New York City. 1890-1913. These are annual averages taken
by permission from E. Frickey, Bank Clearings outside New York City, 1875-1914,
Review of Economic Statistics, vol. VII, 1925, p. 260. They are for seven selected cities.
See Professor Frickey's article for further description of the series.
(8) Output of Manufacturing and Mining Multiplied by Prices. The index of manu-
facturing and mining is described in Chart VI (8). For the wholesale price index see the
Bureau of Labor Statistics Bulletin 284, p. 131 — all-commodity index, 1913 = 100.
(4) Pay Bolls. This curve is based on figures taken by permission from Real Wages
in the United States, 1890-1926, by Paul Douglas, published by Houghton Mifflin Com-
pany, Boston, 1930, pp. 440 and 463. The series is the product of (a) estimated total
numbers of persons employed in manufacturing and transportation ancl (&) average annual
money earnings of persons employed in manufacturing and transportation. For a descrip-
tion of the series the reader is referred to Professor Douglas's book. Since both constituents
are composed of unavoidably rough estimates, the product of the two is clearly a very
questionable indicator of the general course of events, even apart from the fact that it
would in any case not be strictly comparable with the other series used on the chart.
(5) Series (2) Divided by Series (4). Keeping in mind what has been said about Series
(4), this curve is at best a highly conjectural approximation to the measure we should like
to have.
CHART XXIV. PHILADELPHIA CLEARINGS
This chart presents the results of Dr. Georgescu's method applied to Philadelphia Bank
Clearings, 1878-1914. See Chap. V, p. 215.
CHART XXV. UNITED STATES. ANNUAL. LOG SCALE
(1) Deposits outside New York City. The figures are annual averages of the call date
figures of "individual" deposits described above, Chart VI (1).
(2) Clearings outside New York City. See Chart XXIII (2).
(3) Pig iron Consumption. See Chart XII (2).
(4) Equipment Production Index. See Chart XVII (1).
(5) Loans and Discounts outside New York City. These are annual averages of the
loans and discounts of all national banks outside New York City reported for dates of call
from A. A. Young, op. cit., pp. 8-13, see Chart VI (1). Overdrafts are included in loans
and discounts prior to 1898, but not later.
CHART XXVI. UNITED STATES. ANNUAL. LOG SCALE
(1) Expenditure on Producers' Goods. Series (1), Chart XVII multiplied by Series (5),
Chart XL
(2) Expenditure on Consumers' Goods. Series (4), Chart XVII multiplied by Series
(1), Chart XI.
(3) Outside Clearings. See Chart XXIII (2).
(4) Expenditure on Producers' Goods Divided by Outside Clearings. Series (1)
divided by Series (3).
(5) Expenditure on Consumers' Goods Divided by Outside Clearings. Series (2)
divided by Series (8).
CHART XXVII. UNITED KINGDOM. LOG SCALE
(1) Production. See Chart V (2).
(2) Aggregate Money Wage Bill. Annual. From 1860 to 1901 the series is taken from
A. L. Bowley, Tests of National Progress, Economic Journal, vol. XIV, September 1904,
by permission of the author and Messrs. Macmillan & Company, Ltd., London. Professor
APPENDIX 1061
Bowley has since published a revised set of figures; see his recent book on Wages and
Income in the United Kingdom since 1860 (Cambridge University Press, 1937). From
1901 to 1913 the series was taken by permission from A. C. Pigou, Industrial Fluctuations,
2d ed.; Macmillan & Company, Ltd., London, . 1929, pp. 383-384. These figures are
based on Bowley's rates of wages and estimates of variations in the number of the wage
earning population.
(3) Provincial Clearings. 1887-1913. Quarterly averages of monthly data for
Manchester and Birmingham taken by permission from D. S. Thomas, An Index of British
Business Cycles, Journal of the American Statistical Association, vol. XXI, March 1926,
p. 61. The source quoted is the Bankers' Magazine, London.
(4) Production Multiplied by Prices. Series (1) above multiplied by Series (3),
Chart V.
(5) Total Taxable Income, adjusted for changes in the method of assessment. 1842-
1913. Annual. The smoothing effect of taxable income from profits being defined as an
average of actual profits [see below, Chart XXVIII (3)] should be borne in mind. The
items of the series become strictly comparable from 1894 on (160 pounds exemption limit
and repairs allowance) . The series was taken by permission from J. Stamp, British Incomes
and Property, pp. 31&-319, published by P. S. King & Son, Ltd., London, 1916.
CHART XXVIII. UNITED KINGDOM. ANNUAL. LOG SCALE
(1) Wage Kates. The series from 1850 to 1903 has been taken by permission from
G. H. Wood, Real Wages and the Standard of Comfort since 1850, Journal of the Royal
Statistical Society, vol. LXXII, March 1909, pp. 99-103. Mr. Wood's index of money
wages is partly based on Professor Bowley's work and partly on additional material. The
index is a weighted average allowing for changes in the numbers employed in the various
industries (agriculture, building, printing, shipbuilding, engineering, coal, puddling, cotton,
wool and worsted, gas, and furniture). The series has been continued, by means of the
figures of the Board of Trade, by W. T. Layton and G. Crowther, An Introduction to the
Study of Prices, published by Macmillan & Company, Ltd., London, 1936, from which
the relevant items have been taken by permission.
(2) Wages Bill. See Chart XXVII (2).
(3) Profits. Gross Schedule D assessments (Business Profits). In 1908-1909 roughly
three-quarters of the assessments were based on the average of the three preceding years.
As a rough approximation, therefore, the series (except in the cases of canals, railroads,
ironworks, and gasworks, all of which were assessed for profits of the preceding year, and of
mines, which were assessed for the average of the five preceding years, but for which detailed
annual figures are given by Stamp so that individual adjustments were possible) was lagged
by two years and items were plotted for midyear instead of for April. All figures were
taken by permission from J. C. Stamp, op. cit [see above, Chart XXVII (5) ].
(4) Unemployment. See Chart XX (4).
(5) Wage Bill Divided by Wholesale Prices. For wage bill see Chart XXVII (2) . The
figures for wholesale prices were taken by permission from G. F. Warren and F. A. Pearson,
Prices, published by John Wiley & Sons, Inc., New York. Their index joins the Sauerbeck
and Statist indices referred to in Chart IV (3).
(6) Real Wages. Same sources as above, series (1). In estimating real wages Mr.
Wood assumed that four-fifths of total wages were in 1850 spent on commodities other than
housing and that the latter item thence increased steadily (one-half of the increase in rent
being attributed to improving quality of housing and one-half to other causes). For the
rest Mr. Wood took the unweighted average of prices of all commodities of ordinary con-
sumption for which series were then obtainable. The index has been continued on Board
of Trade figures.
106& BUSINESS CYCLES
(7) Real Wages Allowing for Unemployment. Series (6) was corrected by Mr. Wood
(op. cit., Series (1)) by using the trade-union unemployment percentage (Series (4)). From
1896 on this series does not agree with Pigou's [see Chart XX (4)] or that given by the
Abstract of Labour Statistics, which has been used instead.
CHART XXIX. UNITED STATES. LOG SCALE
(1) Wage Rate. 1840-1913. Annual. Index of rate per hour, excluding agriculture,
1910-1914 = 100, during the Civil War on currency basis. It is said to be based on "all
available material." No allowance has been made for the reduction in hours worked per
week. The source is the Monthly Labor Review, vol. XXXII, No. 2, IJebruary 1931, p. 143.
(2) Railroad Earnings. 1866-1913. Monthly figures corrected for seasonal. Gross
earnings of 14 systems as they existed in 1914, made as homogeneous as possible by tracing
back figures for roads later leased or absorbed by these systems. For a further description
see A Monthly Index of Railroad Earnings, 1866-1914 by Arthur H. Cole, Review of
Economic Statistics, vol. XVIII, February 1936. Professor Cole was kind enough to supply
the figures from his files.
(3) Dividends. 1902-1913. Monthly. Payments by industrial corporations, from
the Review of Economic Statistics, Preliminary vol. I, 1919, p. 164. The New York Journal
of Commerce and Commercial Bulletin is the original source.
(4) Wage Bill. See Chart XXIII (4).
(5) Employment in Massachusetts Factories. 1889-1913. The index, 1914 = 100,
was taken by permission from Ralph G. Hurlin, Three Decades of Employment Fluctua-
tions, the Annalist, vol. XVIII (Oct. 24, 1921), pp. 387-388. Obviously, Massachusetts
figures cannot be relied on to depict national employment faithfully.
(6) Commercial Failures. 1857-1913. Aggregate liabilities. From the Statistical
Abstract of the United States for 1914, p. 681.
(7) Wholesale Prices. See Chart IV (1).
CHART XXX. GERMANY. LOG SCALE
(1) Wage Rates. 1850-1913. Pfennigs per man-hour paid to miners in the Ruhr
Valley, Dortmund Mining District, taken by permission from Ernst Wagemann, Economic
Rhythm, translated from the German by D. H. Blelloch and published by the McGraw-Hill
Book Company, Inc., New York, p. 265.
(2) Wages Bill. Saxony. The figures were taken by permission from the Vierteljahrs-
hefte zur Konjunkturforschung, Erganzungsheft vol. II No. 3, 1927, p. 33, edited by Pro-
fessor Wagemann. Conditions in the Saxon Kingdom were considered as sufficiently
typical of German conditions in general.
(3) Wholesale Prices. 1850-1913. The figures were taken from Wirtschaft und
Statistik, vol. V Sonderheft 1, 1925, p. 19, and the Statistisches Jahrbuch, vol. XLVII, 1926,
p. 263. Until 1878, the source is Adolf Soetbeer, Materialien zur Erlauterung und Beur-
teilung der Wirtschaftlichen Edelmetallverhaltnisse und der Wahrungsfrage, Berlin, 1886.
(4) Profits in Saxony. From the article quoted in Series (2).
(5) Unemployment. 1906-1914. End of month figures of industrial unemployment
taken by permission from Ernst Wagemann, Konjunkturlehre, published by R. Robbing,
Berlin, 1928, p. 195.
(6) Series (1) Divided by Series (3).
CHART XXXI. UNITED KINGDOM. ARITHMETIC SCALE
(1) Real Wages, Allowing for Unemployment. See Chart XXVIII (7).
(2) Nine-year Moving Average of (1).
(3) Deviations from nine-year moving average.
APPENDIX 1063
CHART XXXII. GERMANY. ANNUAL. LOG SCALE
(1) Loans and Discounts. Nine Berlin Grossbanken. End of year figures.
(2) Total Deposits (in the English sense of the term). Nine Berlin Grossbanken.
End of year figures.
(8) Note Circulation of all German Note-issuing Banks.
All three series are based on the official figures. For a full presentation and discussion
of these and related data see L. A. Hahn, Zur Frage des volkswirtschaftlichen Erkennt-
nisgehalts der Bankbilanzziffern, Vierteljahrshefte zur Konjunkturforschung, vol. I, 1926,
Erganzungsheft 4.
CHART XXXIII. UNITED STATES. LOG SCALE
(1) Individual Deposits plus Circulation outside New York City. See Chart VI (1).
(2) Loans and Discounts outside New York City. See Chart XXV (5).
(3) Individual Deposits plus Circulation minus Investments outside New York City.
Series (1) minus the investment series described in Chart XXIII (1).
Actual call date figures have been used in all three cases.
CHART XXXIV. UNITED STATES. LOG SCALE
(1) Production of Industrial Equipment. See Chart XVII (1).
(2) Building Permits. See Chart XIV (5).
(8) Investments. National Banks outside New York City — a somewhat
sample. See Chart XXIII (1). A five-item moving total was plotted.
(4) New Security Listings. Stocks and bonds listed on the New York Stock Exchange
minus old issues and issues replacing existing securities. The series was taken from the
Commercial and Financial Chronicle, by permission of the William B. Dana Company,
New York, and from the Financial Review.
(5) Loans and Discounts. National Banks outside New York City. See Chart XXV
(5). A five-item moving total was plotted.
CHART XXXV. BANK OF ENGLAND FIGURES, ETC. 1844-1914.
LOG SCALE
All series except (2) and (3) are annual averages taken by permission from R. H. I.
Palgrave, Bank Rate and the Money Market, published by John Murray, London, pp.
12-13, for the years up to 1900, and from the Bankers' Almanac and Yearbook, 1935-1936,
for the years 1900-1914.
(1) Private Deposits at the Bank of England. There is a break in 1873, owing to the
fact that since then Chancery Balances have been put under the head of government
deposits. The average amount of these balances was 1 million pounds.
(2) London Total Clearings Divided by Total English Deposits. The clearings are
monthly average clearings per working day during June or December multiplied by 306,
taken by permission of the University of Chicago Press, Chicago, from The Velocity of
Bank Deposits in England, by Lionel D. Edie and Donald Weaver, Journal of Political
Economy, vol. XXXVIII, No. 4, August 1930, p. 897. The deposits are monthly average
total deposits in the joint stock and private banks of England and Wales (except the Bank
of England) on June 30 or Dec. 31, taken by Edie and Weaver from the Economist.
(3) London Total Clearings See Series (2).
(4) Proportion of Reserve (Banking Department) to Deposits plus Bank Post Bills.
(5) Proportion of London (Clearing) Bankers' Balances to the Reserve of the Banking
Department of the Bank of England,
1064 BUSINESS CYCLES
(6) London Bankers' Balances at the Bank of England.
(7) "Other Securities" in the banking department.
(8) Notes Held by the Public.
CHART XXXVI. UNITED STATES. LOG SCALE
(1) Bank Clearings, New York City. 1866-1914. From 1866 to 1902 the series was
taken from Ada Matthews, New York Bank Clearings and Stock Prices, 1866-1914, Review
of Economic Statistics, vol. VIII, 1926, p. 188. The data were used by permission of the
Commercial and Financial Chronicle, which is the source quoted by MissJMatthews. From
1908 to 1914 the data were taken from the Wall Street Journal by permission of Dow, Jones
& Co., Inc., New York. Four-quarter moving totals of quarterly totals centered on the
third item have been plotted.
(2) Value of Transactions, New York Stock Exchange. 1875-1914. Annual. Num-
ber of shares traded multiplied by average price. The figures for 1875 to 1909 were taken
from the United States National Monetary Commission, Statistics for the United States,
1867-1909, p. 9, where the Commercial and Financial Chronicle is given as the source, while
the figures for 1909 to 1914 came directly from the Commercial and Financial Chronicle.
(8) Call Loan Rate, New York Stock Exchange. 1866-1914. Moving 12-month
totals (centered on the seventh month) of monthly averages of daily renewal rates. The
series was taken by permission from the Standard Trade and Securities Service, Standard
Statistical Bulletin, Base Book, January 1982, p. 42, and April 1934, p. 6. The sources
quoted there are Ogle, Dunn & Company, whose permission to use their material is herewith
acknowledged, the Review of Economic Statistics, and the Financial Review.
(4) Loans and Discounts. New York City. 1870-1914. See Chart XXV (5). Five-
item moving totals of call date figures have been plotted.
(5) Deposits. New York City. 1870-1914. " Net " deposits of national banks. For
definition of "net" deposits and also for the source see Chart XXIII (1). Five-item mov-
ing totals of call date figures have been plotted.
(6) Railroad Stock Prices. Monthly. From 1854 to 1882 the series was taken by
permission from the Annalist, vol. XL, Oct. 28, 1932), p. 580. It is the Clement Burgess
index of stock prices, average of high and low figures, adjusted for stock dividends. From
1888 to 1935 the index of 20 railroads — weighted by the number of shares of stock out-
standing— of the Standard Trade and Securities Service, Standard Statistical Bulletin,
April 1934, p. 30, has been used by permission. The two series were spliced together.
(7) Industrial Stock Prices. Monthly. From 1883 to 1900 the Axe-Houghton
weighted average of 10 industrial stocks has been used by permission of the Annalist,
vol. XXXVII, Jan. 16, 1931, p. 177. From 1900 to 1913 the index used was taken by
permission from the Standard Trade and Securities Service, Standard Statistical Bulletin,
April 1934, p. 30. It is composed of the prices of 50 industrial stocks weighted by value of
shares outstanding, corrected for rights, stock dividends, changes in par value, and con-
solidations. Both indices are averages of high and low. They were spliced together.
CHART XXXVII. GREAT BRITAIN. LOG SCALE
(1) British Stock Prices. 1867-1914. End of month data; 1890 = 100; unweighted
arithmetic average. The index was taken by permission from the London and Cambridge
Economic Service, Special Memo. 87, An Index Number of Securities, 1867-1914, by K. C.
Smith and F. C. Home. The stocks included divide up into the following divisions:
(1) coal, iron, etc; (2) electrical goods; (3) textiles; (4) food; (5) drink; (6) building
materials; (7) lighting; (8) chemicals; (9) stores; (10) miscellaneous; (11) transport and
communication.
APPENDIX 1065
(2) London Call Rate. 1888-1914. Annual average rate on money at call or short
notice taken until 1906 from the United States National Monetary Commission's Statistics
for Great Britain, Germany, and France, 1867-1909, p. 143. From 1907 to 1914 the
figures are annual averages of rates on floating money as published by the Economist.
(3) London Clearings. 1870-1914. Total amounts cleared from 1870 to 1902; after
that, "town clearings." Both series are from the Statistical Abstract for the United Kingdom
and were spliced together.
CHART XXXVIII. GERMANY. LOG SCALE
(1) Stock Prices. 1870-1913. Monthly. This series was taken, by permission, from
Otto Donner, Die Kursbildung am Aktienmarkt, Vierteljahrshefte zur Konjunkturforschung,
Sonderheft 36, 1934, p. 98. It is composed of quotations for an increasing number of com-
panies, 70 of which were available as far back as 1890, and includes banks, shipping com-
panies, railways, mines, electric companies, manufacturing, and building.
(2) Industrial Bond Issues. 1883-1913. Annual. Domestic issues (market value)
from H. Kleiner, Die Emissions-Statistik in Deutschland, Miinchner Volkswirtschaftliche
Studien 131, 1914, Table I, pp. 119-124. The main source of the material embodied in that
table was Der Deutsche Okonomist. The data compiled by this periodical refer to listings at
the stock exchanges rather than to issues and are not quite complete, especially as regards
the listings at the smaller stock exchanges. In these and other respects other compila-
tions, such as those of the Frankfurter Zeitung, the Imperial Statistical office (since 1897),
or the statistical department of the Reichsbank may be preferable. But the material of the
Deutsche Okonomist yields the longest homogeneous series, and on the whole it seemed best
to use it.
(3) Industrial Stock Issues. From the same source as (2).
(4) Foundations. 1871-1913. Annual. " Nominal" capital of newly founded joint
stock companies, from Der Deutsche Okonomist, vol. XXVI, 1908, p. 28, and vol. XXXII,
1914, p. 412.
(5) Dividends, Per Cent of "Nominal" Capital. 1870-1913. Annual. This series
has been taken, by permission, from the study by Otto Donner, see above, Series (1).
Data are for the industrial companies included in the stock price index of the Statistisches
Reichsamt. From 1890 on, the percentages are weighted according to the share capital
of the various companies.
CHART XXXIX. UNITED STATES POSTWAR PULSE. LOG SCALE
(1) Deposits plus Circulation. The deposits are net demand deposits of weekly report-
ing member banks in 100 leading cities outside New York. The monthly average figures
used (as revised in 1929) were supplied to the author from the files of the Harvard Economic
Society (now Committee on Research in the Trade Cycle) This series, subsequently
reported through Aug. 23, 1935, was at the time of writing available only through February
1933. Therefore, it was spliced to a series which covered 89 cities from January 1932, and
90 cities from January 1934. Since the Banking Act of 1935 changed the definition of net
demand deposits, it should be observed that that term is being used in its old sense here:
it includes all demand deposits minus United States government deposits, but balances due
from banks and trust companies subject to immediate withdrawal and cash items in process
of collection were (for each bank separately) deducted from balances due to other banks
and subject to immediate withdrawal, so that it may still be said that, very roughly, these
figures indicate the variations in the demand deposits held by the "public."
The series of money in circulation was taken, by permission, from J. W. Angell, The
Behavior of Money, pp. 178-179, published by the McGraw-Hill Book Company, Inc.,
New York, 1936. It represents a monthly estimate of currency issued and not yet redeemed
1066 BUSINESS CYCLES
minus currency held in the Treasury (as asset), in federal reserve banks or with federal
reserve agents, and vault cash of all reporting banks (as reported to the comptroller of the
currency). The figures for vault cash are available only for June 30 of each year, and had
to be estimated from call date information for all member banks' vault cash. Monthly
data exist, however, for the sum of currency in outside circulation plus vault cash of all
banks. Since addition of those figures for deposits and circulation would have given much
too great a weight to circulation, the proportion was calculated of net demand deposits of
the above banks to total "circulating deposits" as estimated by Angell, op. tit. It turned
out to be fairly steady, the average for the year 1926 being 87.6 per cent. On obviously
very simplifying and bold assumptions, therefore, 38 per cent of that circulation was added
to those net demand deposits of weekly reporting member banks. It is impossible to use
Angell's "circulating deposits," since they include New York City and also United States
deposits.
(2) Interest Rate. Monthly. From 1919 to 1931, New York prime commercial paper
rate, taken by permission from the Standard Trade and Securities Service, Standard
Statistical Bulletin, Base Book, January 1932. From 1932 to 1934, the rate on four to six
months' paper has been taken from the Survey of Current Business.
(3) Production Index. Monthly. This is the Federal Reserve Board's seasonally
adjusted index of manufacturing and mining, taken from the Federal Reserve Bulletin,
vol. XIX, September 1933, p. 584, and vol. XXI, May 1935, p. 282. It is a weighted
average per working day. The 1923-1925 average value was used for weighting in the
case of mining, and the 1923 value added by manufacture in the case of manufacturing.
For a further description of the index, see the Federal Reserve Bulletins, vol. XIII, No. 2,
February 1927; vol. XIII, No. 3, March 1927; and vol. XVIII, No. 3, March 1932.
(4) Wholesale Prices. Monthly. This is the Bureau of Labor Statistics index on the
base 1910-1914 = 100. The new index has been used for the years 1932-1934 by shifting
its base (1926) to the base 1910-1914.
CHART XL. GERMAN POSTWAR PULSE, LOG SCALE
(1) Wholesale Prices. 1919-1934. Monthly averages. This is the index of the
Statistisches Reichsamt, taken from Wirtschaft und Statistic (1913 = 100). From 1919 to
1924 this is an arithmetic average of prices of less than 50 commodities, essentially raw
materials and semi-manufactured goods, weighted according to prewar consumption.
From 1924 to 1934 it is an arithmetic average of about 400 commodities (including finished
products) on the basis of quotations increasing from 800 to 1,000, weighted by an average of
prewar and postwar consumption (where postwar consumption approaches prewar con-
sumption) or on 1925 consumption figures. Groups are weighted as follows: agricultural
commodities 35 per cent, imported groceries 3 per cent, industrial raw materials and semi-
manufactured products 88 per cent, manufactured goods 24 per cent. The two indices
were spliced together. For more detailed information see Wirtschaft und Statistik, vol. VI,
1926, p. 875, and Vierteljahrshefte zur Statistik des Deutschen Reiches, vol. XXXVI, 1927,
p. 37, and vol. XLI, 1932, p. 139.
(2) Interest Rate. Rate on bank acceptances, monthly average. Taken by permis-
sion from the Institut fiir Konjunkturforschung, Konjunktur-Statistisches Handbuch, 1936,
p. 113.
(3) Production. Quarterly index of industrial production per working day (1928 ==
100), taken by permission from the Institut fttr Konjunkturforschung, Konjunktur-Statis-
tisches Handbuch, 1936, p. 52. Seasonal variations are eliminated. The Saar is included
from March 1935, so that the series is not strictly homogeneous. The index was first
calculated in 1927, and revised in 1929, 1931, 1933, and 1935. It is an arithmetic average.
Series are weighted within each group by value added, number employed, and horse-power
APPENDIX 1067
installed. Groups are separately weighted by value added. The index now represents
66 per cent of industrial net production. For further discussion, see Vierteljahrshefte zur
Konjunkturforschung, vol. IV, No. 4A, 1930; vol. VI, No. 1A, 1931; vol. VII, No. 4A, 1933;
and the Wochenbericht of the Institut filr Konjunkturforschung, vol. VIII, No. 24, for
June 19, 1935.
(4) Deposits and Circulation. The deposits (Glaubiger, mostly what in this country
would be called business deposits) are those of five big banks (evidently taken as a sample,
but an unsatisfactory one, of total deposits), end of month figures from the Institut fUr
Konjunkturforschung, Konjunktur-Statistisches Handbuch, 1936, p. 136. The five banks
are: Deutsche Bank und Disconto-Gesellschaft, Berlin; Dresdqer Bank, Berlin; Commerz-
undPrivat Bank, Berlin; Bayerische Hypotheken-und Wechsel-Bank, Mlinchen; Allgemeine
Deutsche Credit-Anstalt, Leipzig. Figures include the effects of amalgamations. No
figures are available for January of each year, and from 1925 to 1927 figures are given only
for February, April, June, August, October, and December. The circulation series is total
money in circulation, end of month figures from the Institut fiir Konjunkturforschung,
Konjunktur-Statistisches Handbuch, 1936, p. 130. No allowance is made for money in
banks. *
CHART XLI. BRITISH POSTWAR PULSE. LOG SCALE
(1) Wholesale Prices. 1919-1934. Monthly. The index, 1913 = 100, was taken
from the Board of Trade Journal. For a description see A. W. Flux, The Measurement of
Price Changes, Journal of the Royal Statistical Society, vol. LXXXIV, Part 2, March 1921.
It is the geometric average of the wholesale prices of 150 commodities divided into eight
groups of approximately equal total values of output or, in the case of imported consumers'
goods, of imports in 1907. The basis of valuation is the census of production of 1907.
Within the eight groups, classes of commodities are represented by a number of price series
varying according to the same criterion. The newly revised index covering 200 commodi-
ties and weighting according to the values of the 1930 census of production presents almost
exactly the same picture from 1930 to 1934. Therefore, it was not thought necessary to
redraw the curve when the new index became available.
(2) Interest Rate. Monthly. Average rate on three months' commercial paper for
week ending the fifteenth of the month. Data were taken by permission from the monthly
bulletins of the London and Cambridge Economic Service.
(3) Deposits plus Circulation. The deposits are current accounts of the 10 London
Clearing Banks, average for the month, taken from the Report of the Committee on Finance
and Industry, 1931, Cmd. 3897, pp. 284-289, and for later years from the Bank of England
Statistical Summary. In the case of the National Bank, Ltd., only figures relating to
offices in England are included. Some items in the earlier years are estimates. The
circulation series represents Bank of England and currency notes in circulation from the
eleventh to the seventeenth of each month, taken, by permission, from the monthly
bulletins of the London and Cambridge Economic Service.
(4) Production Index. Quarterly index of production taken, by permission, from the
London and Cambridge Economic Service, Monthly Bulletin, vol. VIII, No. 4 (Special
Quarterly Issue, Apr. 23, 1929) and later Special Quarterly Issues. This is an arithmetic
average, 1924 = 100, of individual output series, weighted according to net output as given
by the 1924 census of production. For further discussion see the Special Memorandum 8
of the London and Cambridge Economic Service, The Physical Volume of Production, by
J. W. F. Rowe. The subindexes are (1) coal mining; (2) pig iron, steel, shipbuilding,
railroad vehicles; (3) copper, lead, tin, zinc; (4) cotton, silk; (5) wheat, flour, cocoa, tobacco;
(6) oil seed crushings, heavy chemicals; (7) paper.
1068 BUSINESS CYCLES
CHART XLII. INDUSTRIAL PRODUCTION. ANNUAL. LOG SCALE
(1) United States. Index taken, by permission, from the Standard Trade and Securi-
ties Service, Standard Statistical Bulletin, December 1935, p. 39. A weighted composite of
64 series. Corrected for seasonal. Revised in 1933. For series and weights see the
Standard Trade and Securities Service, Basic Statistics, vol. LXXX, No. 29, June 5, 1936,
p. D-36.
(2) Germany. See Chart VII (2).
(3) Great Britain. Hoffmann Index. See Chart V (2).
(4) Great Britain. London and Cambridge Economic Service Index. See Chart XLI
(4).
CHART XLIII. UNITED STATES PRODUCTION SERIES. 1919-1934.
MONTHLY. LOG SCALE
All the series taken by permission from Y. S. Leong, Indexes of the Physical Volume of
Production of Producers' Goods, Consumers' Goods, Durable Goods, and Transient Goods,
Journal of the American Statistical Association, vol. XXX, No. 189, June 1935. All series
are on a daily average output basis and adjusted for seasonal variation. The 1923-1925
average = 100. The aggregative method was used to combine the individual series
(groupwise) into composite index numbers. Weighting is by value added by manufacture.
The average value added of the census years 1923, 1925, and 1927 has been used for
producers' goods, all durable goods, and total manufacturing output from 1922, and for
consumers' goods excluding motorcars, and for transient goods from 1923 on. Those
weights seemed, however, inappropriate for the preceding years. Accordingly, Mr. Leong
computed another set of index numbers, using the value added figures of 1919, and com-
bined it with the first set. The resulting index is, for 1919 to 1922, a simple geometric
average of the two in the case of consumers' goods excluding motorcars, and of transient
goods, and for 1919 to 1921 a geometric average with variable weights in the case of
producers' goods, all durable goods, and total manufacturing output. This procedure,
though not easy to defend on general principles, probably represents in this case a fair
approximation to the theory outlined in Chap. IX.
(1) Producers' Goods. Unfinished goods or goods used to produce other goods.
Certain textiles, forest products, paper and printing, chemicals, leather, stone and clay, iron
and steel, nonferrous metals and their products, transportation equipment.
(2) Consumers' Goods. Series representing fabricated goods for immediate or nearly
immediate consumption. Food and kindred products, certain kinds of textiles and print-
ing, gasoline and kerosene, rubber products, shoes, gloves, radiators, sanitary ware, auto-
mobiles, tobacco manufactures.
(3) Consumers' Goods Excluding Automobiles. Same as Series (2), except for
automobiles.
(4) Durable Goods. Goods with an average useful life of more than two years. Forest
products, coke, stone, clay, glass, iron, steel, nonferrous metals and their products, and
transportation equipment.
(5) Transient Goods. Products with an average useful life not exceeding two years.
Food and kindred products, textiles and their products, paper and printing, chemical,
rubber, and leather products.
(6) Manufactures. All the series used in the above indices: 15 food and kindred prod-
ucts; 4 textiles and textile products; 2 forest products; 10 paper and printing; 7 chemical
and allied products; 2 rubber products; 5 leather and leather products; 2 stone, clay, and
glass products; 4 iron and steel products; 4 nonferrous metals and their products; 5 trans-
portation equipment; 3 tobacco manufactures. These items represent directly about
APPENDIX 1069
50 per cent of the value added by manufacture in all manufacturing industries in the census
years 1923, 1925, and 1927.
CHART XLIV. UNITED STATES CUSTOMERS' LOAN RATES. ANNUAL.
ARITHMETIC SCALE
All three series are yearly averages of monthly averages as reported by banks to the
Federal Reserve Board. The monthly averages are based on rates reported for three types
of customers' loans, commercial loans, and time and demand loans on securities. The
averages are weighted according to the relative importance of each of these three types of
loans, the relative importance of each reporting bank, and the relative importance of each
city in each group.
CHART XLV. UNITED STATES PROFIT RATIOS. ANNUAL.
ARITHMETIC SCALE
All series were taken, by permission, from W. L. Crum, Corporate Earning Power in the
Current Depression, Business Research Studies 10, Harvard University Graduate School of
Business Administration. The profit ratio is statutory net income (minus federal taxes),
divided by gross income. The data are from federal income tax reports. For further
explanation, see the article mentioned.
CHART XLVL UNITED STATES. 191&-1934. MONTHLY. LOG SCALE
(1) Outside Debits. Bank debits to individual deposit accounts in 140 cities outside
New York City as published by the Federal Reserve Board. For the most part, these
debits arise from checks against depositors' accounts and represent payments. All debits
to deposit accounts of individuals, firms, corporations, and U. S. government, county, and
municipal accounts enter the total, however, including debits to war loan deposit accounts,
savings accounts, payments from trust accounts, and certificates of deposit paid, except
debits in settlement of clearinghouse balances, debits to accounts of other banks, payments
of cashiers' checks, charges to expenses and miscellaneous accounts, corrections, and the
like. See the Twenty-Second Annual Report of the Federal Reserve Board, 1935, p. 175.
(2) Outside Net Demand plus Time Deposits minus Outside Investments. See
Chart L (4) and (7).
(3) Index of Pay Rolls in Manufacturing Industry. This is the revised index of the
Bureau of Labor Statistics, see the Bureau's Bulletin 610, p. 22. It represents average
weekly factory pay rolls (1923-1925 = 100). In 1925 54 industries were covered, employ-
ing 83 per cent of the workers in all manufacturing industries. The establishments surveyed
employed about 50 per cent of the 83 per cent. Since 1931 coverage extends to 90 indus-
tries. Weighting is according to pay rolls in 1923-1925. The index is not corrected for
seasonal variations. Another correction has, however, been made which should be men-
tioned. Since the index revealed a downward bias when compared with the trend shown by
the figures of the biennial census, it was adjusted so as to conform to that trend. The
operation consisted essentially in straight-line adjustments to the averages of pay rolls for
pairs of census years: an appropriate cumulative unit was applied to the 24-month interval,
in order to bring up the average for each year to the amount indicated by the census figure.
For example, the original data for the years 1919 to 1921 were thus adjusted to the annual
census averages for those two years. Then, before going on to the adjustment to the census
figures from 1921 to 1923, preliminary adjustment for 1921 to 1923 became necessary in
order to reestablish the comparability destroyed by the previous adjustment.
(4) Outside Net Demand Deposits. See Chart XXXIX (1), first component.
(5) Annual Aggregate Realized Income. Morris A. Copeland, in How Large is Our
National Income?, Journal of Political Economy, vol XL, December 1932, p. 773, presents
1070 BUSINESS CYCLES
a series labeled "realized income revised," which is a revision of Wilford King's series. It
excludes both undistributed profits and capital gains. W. L. Crum, in turn, revised Pro-
fessor Copeland's series, see The National Income and its Distribution, Journal of the
American Statistical Association* vol. XXXI, March 1935, p. 36, deducting "imputed
income yielded by individually owned urban homes and other durable consumers' goods"
by means of the figures given in America's Capacity to Consume, by Maurice Leven, H. G.
Moulton, and Clark Warburton, Brookings Institution, 1934, p. 153.
For the chart Professor Crum's revised series, taken by permission from the Journal of
the American Statistical Association was spliced on to the series prepared by the Depart-
ment of Commerce as published in the Survey of Current Business, July, 1936, by Robert R.
Nathan, Chief, Income Section, Division of Economic Research. This is a continuation
of the work presented in National Income 1929-1932, Senate Document 124, Seventy-third
Congress, Second Session. The figures used are those of "national income paid out."
They represent payments to, or receipts by, individuals, such as wages, salaries, interest,
dividends, entrepreneurial withdrawals, net rents, and royalties, and differ from "income
produced" by positive or negative business savings.
(6) Outside Debits Divided by Outside Net Demand plus Time Deposits. See Series
(1) and Chart L (4).
(7) Hourly Earnings in Manufacturing. This series is labeled wage rates on the chart
but, of course, may vary in responses to changes in the amount of overtime and in the
composition of the working force. The figures are quarterly for 24 industries, from 1920 to
1926, and monthly for 25 industries, from 1927 on. Data through 1926 were taken by
permission from the National Industrial Conference Board, Wages in the United States,
1914-1930, p. 44, and from 1927 to date from the Survey of Current Business, vol. XII,
No. 12, December 1932, p. 18, and later numbers. At the end of 1931 over 1,400 plants
were covered. The weekly earnings from which the hourly rates are derived are computed
by weighting the average weekly earnings in each industry by the relative importance of the
industry as revealed by the Census of Manufactures of 1923. Moreover, the weights are
made to reflect the relative importance of each labor group in each industry as
ascertained by the Conference Board's investigations during 1927—1929. Data on hours of
work are from the same number of plants and workers as the weekly earnings. The base
is 1923 = 100.
This chart was made up before the National Industrial Conference Board's volume on
Wages, Hours, and Employment in the United States, 1914-1936, became available. That
volume carries the series back all the way on a monthly basis and, while the general method
is the same, certain revisions in the details of procedure have been made. Results differ,
however, so insignificantly from the series above described that it was not thought necessary
to redraw the chart and to recalculate the series into which these figures enter [Chart XL VII
(6) and (7)].
CHART XLVII. UNITED STATES. LOG SCALE
(1) Hourly Wage Rates. See Chart XLVI (7).
(2) Pay Rolls. See Chart XLVI (3).
(3) Employment. Manufacturing Industries. For source and description of the
material used see Chart XLVI (3) [Pay rolls]. In this index the weights are based on the
annual average number of wage earners employed in the industry or group from 1923 to
1925.
(4) Cost of Living. This index has been constructed by the National Industrial Con-
ference Board, and is here used by permission. It aims specifically at the prices relevant
to the American wage earner's budget (1923 = 100). The five major groups (postwar
budget weights) are food (33 per cent), housing (20 per cent), clothing (12 per cent), fuel
APPENDIX 1071
and lighting (5 per cent), and sundries (30 per cent). The food index used is the Bureau
of Labor Statistics retail food price index for the fifteenth of each month. From 1920 to
1925 the figures are based on three comprehensive surveys for each March, July, and
November, and less comprehensive information for the other months. Since 1925, monthly
calculations have been made on a comprehensive basis. The figures were taken from the
Standard Trade and Securities Service, Standard Statistical Bulletins, and the U. S. Depart-
ment of Commerce, Survey of Current Business, vol. XVI, No. 1, January 1936, p. 19. An
improved index, which differs from the one used sometimes by as much as 2 per cent, has
been published since in the National Industrial Conference Board's volume on Cost of
Living in the United States, 1914-1936. It gives somewhat lower figures for the earlier
and somewhat higher figures for the later years.
(5) Wholesale Prices. See Chart XXXIX (4).
(6) Series (1) Divided by Series (5).
(7) Series (1) Divided by Series (4).
(8) Series (2) Divided by Series (4).
(9) Series (2) Divided by Series (5).
CHART XLVIII. GERMANY. LOG SCALE
(1) Unemployment. The number of unemployed registered at the employment offices.
End of month figures. Data taken from the Institut ftlr Konjunkturforschung,
Konjunktur-Statistisches Handbuch, 1936. From March 1935 the Saar country is included.
(2) Wage Rates. Average standard hourly (or piece) rates of workers of the highest
standard age group. First of the month figures, taken from the Institut filr Konjunktur-
forschung, Konjunktur-Statistisches Handbuch, 1936. The old index of the statistical office
of the Reich from 1925 to 1927 was chained to the new index, which goes back to 1928.
The old index included a smaller number of industrial groups and only male workers.
The new index includes female workers and is weighted according to the average number
employed in the years 1928 to 1930. For further information sec the Vierteljahrshefte zur
Statistik des Deutschen Reichs, vol. XL, 1931, Heft 2, pp. 94 et seq.
From 1932 to 1934 a third index has been used. It represents standard rates in the
highest age group of unskilled male workers, and was taken from the Statistisches Jahrbuch
des Deutschen Reichs, vol. LV, 1934, p. 279. It was spliced on to the other index.
Of course, standard wage rates cannot be trusted to represent the price of labor accu-
rately.
(3) Wage Bill. Income from wages and salaries according to the estimate of the Insti-
tut fur Konjunkturforschung — excluding pensions and including the income from all
emergency and relief employment. The items for the second and third quarter of 1935
do not include the Saar. The figures were taken from the Institut flir Konjunkturforschung,
Konjunktur-Statistisches Handbuch, 1936.
(4) Dividends. The data of the Institut fiir Konjunkturforschung's Konjunktur-
Statistisches Handbuch, 1936, are believed to give a rough picture of the course of total
dividends. Two figures are presented for each quarter. One is comparable, as regards
companies included, with the figure given for that quarter of the preceding year, and the
other is comparable with the figure for the corresponding quarter of the following year.
Link relatives could not be constructed between the different quarters in the same year.
Therefore, the fourth quarter of 1926 was used as the basis of link relatives of the items for
the fourth quarter of each year, because more than half of all dividends are paid out in the
fourth quarter. However, any significant change in the dates of dividend payments
would destroy any value the above-described relatives may have as an index of total
dividends paid out.
1072 BUSINESS CYCLES
(5) Cost of Living. Both the old index of the Statistisches Reichsamt covering the
years from 1925 to 1928 and the new index constructed from 1928 on were taken from the
Institut ftir Konjunkturforschung, Konjunktur-Statistisches Handbuch, 1936. From 1925
to 1928 groups are weighted mostly on the basis of the budget inquiry of 1907, but some
account was taken of the changes which occurred since the war. From 1928 on, the index is
weighted according to the results of the budget inquiry of 1927-1928. The old and new
indices were linked together by the chain method.
(6) Wholesale Prices. See Chart XL (1).
(7) Series (2) Divided by Series (5).
(8) Series (2) Divided by Series (6).
(9) Series (3) Divided by Series (6).
(10) Series (3) Divided by Series (5).
CHART XLIX. GREAT BRITAIN. LOG SCALE
(1) Profits. 1920-1934. Annual. General index of the returns to "industrial capi-
tal" as a whole. Base 1924 = 100. No allowance was made for the increase in capital
invested since 1924. The "capital" includes debentures and other items, the returns on
which vary slowly. The data for 1932, 1933, and 1934 are provisional. The figures were
taken from Sir J. C. Stamp's letter to the London Times, Aug. 9, 1935, p. 6. Obviously
this series indicates but very imperfectly the variations we would like to measure.
(2) Wage Rates. The figures were taken by permission from A. L. Bowley, A New
Index Number of Wages, London and Cambridge Economic Service, Special Memo 28.
From 1919 to 1925 it is an unweighted index of wage rates in 11 occupations (December
1924 = 100). This was spliced on to the new index of average weekly wages, which is
available back to 1925. The new index covers 20 occupations and is weighted according
to the wage bill of each group in the base period. Figures are for the fifteenth of the month.
Professor Bowley has since revised his figures. For the new index, see again his recent
book, Wages and Income in the United Kingdom since 1860.
(3) Employment. This index was taken from the supplements to the Economist,
Oct. 21, 1933, p. 6, and later numbers. It is based on the number of insured workers in
employment in Great Britain, as published in The Ministry of Labour Gazette for the third
week of each month. Seasonal fluctuations were not eliminated (1924 = 100).
(4) Commercial Failures. The annual number of commercial failures in England and
Wales, taken from the Statistical Abstracts for the United Kingdom.
(5) Wholesale Prices. See Chart XLI (1).
(6) Cost of Living. This index is an arithmetic average (July 1914 = 100). Retail
price quotations (first of the month) are obtained from large and small towns, and for a
number of articles in each of the five major groups. Each group is weighted according
to its importance in the budget of a "typical" workingman's household as ascertained by
the Board of Trade Study of 2,000 workingmens' families made in 1904. The index has
been taken from the Abstract of Labour Statistics. See The Cost of Living Index Number,
Method of Compilation, H.M.S.O.
(7) Series (2) Divided by Series (5).
(8) Series (2) Divided by Series (6).
CHART L. UNITED STATES. 1919-1934. MONTHLY. LOG SCALE
(1) Net Demand Deposits, New York City. Monthly averages of weekly data of
weekly reporting member banks in New York City. Figures as revised in 1929. Federal
Reserve Board data. See Series (2).
(2) Total Deposits — Time and Demand: New York and Outside Net Demand plus
Time Deposits of Weekly Reporting Member Banks in Leading Cities. Monthly aver-
APPENDIX 1073
ages of weekly figures. Data as revised in 1929 (see Federal Reserve Bulletin, vol. XV,
No. 1, January 1929). For further description of net demands deposits see Chart XXXIX
(1), first component. Remarks on the changing number of cities and splicing operations
also apply to Time Deposits.
(3) Bank Debits outside New York City. See Chart XLVI (3).
(4) Outside Time plus Net Demand Deposits. See Series (2) above.
(5) Loans and Discounts of Reporting Member Banks outside New York City. For
changing number of cities and splicing see Chart XXXIX (1), first component. Monthly
averages as revised 1929. For further description of the series, see the Federal Reserve
Bulletin, vol. XV, No. 1, January 1929.
(6) Net Demand Deposits outside New York City. See Chart XXXIX (1).
(7) Outside Investments. Monthly averages of figures for weekly reporting banks
outside New York City. Federal Reserve Board data as revised in 1929. For number of
cities and splicing operations see Chart XXXIX (1), first component. For further descrip-
tion see the Federal Reserve Bulletin, vol. XV, No. 1, January 1929.
CHART LI. LONDON CLEARING BANKS. LOG SCALE
(1) Commercial Bills Discounted. Ten London Clearing Banks. For source see
Chart XLI (3).
(2) Treasury Bills Discounted. See Series (1).
(3) Treasury plus Commercial Bills Discounted. 1930-1936. These are figures for
nine London Clearing Banks (excluding the National Bank, Ltd.). Monthly averages,
taken by permission from the monthly bulletins of the London and Cambridge Economic
Service.
(4) Advances (Loans and Overdrafts). From January 1919 to March 1931, see Series
(1). The figures for nine Clearing Banks (see Series (3)) were spliced on for 1931-1936.
(5) Investments. From 1919 to 1931 see Series (1). Monthly averages for the nine
Clearing Banks, see Series (3), were spliced on from 1931 to 1936. Investments are exclu-
sive of investments in affiliated banks.
(6) Deposit (Time) Accounts. From 1919 to 1931, see Series (1). From March 1931
on, the figures were taken from the Bank of England Statistical Summary. There were
minor changes in 1931 in the allocation of certain deposits. The 1936 figures are for 11
banks.
(7) Current (Demand) Accounts. Same sources as Series (6).
(8) Country plus Provincial Clearings. These are country clearings at the London
clearinghouse plus clearings at 11 provincial clearinghouses. From 1920 to 1928 these
figures were taken by permission of the University of Chicago Press, Chicago, from L. D.
Edie and D. Weaver, Velocity of Bank Deposits in England, Journal of Political Economy,
vol. XXXVIII, No. 4, August 1930. For the rest of the period this series has been spliced
on to the one published by the London and Cambridge Economic Service.
(9) Country plus Provincial Clearings Divided by Current Accounts of Ten London
Clearing Banks. 1920-1928. The clearings are those described in Series (8) and the
current accounts are as estimated by L. Edie and D. Weaver, op. cit., p. 395. The reader
is referred to that article for further discussion of the series and the method of computation.
CHART LII. UNITED STATES. LOG SCALE
(1) Call Loan Rate, New York Stock Exchange. Monthly averages of daily renewal
rates. The series was taken, by permission, from the Standard Trade and Securities
Service, Standard Statistical Bulletin, Base Book, January 1932 and later numbers. The
sources quoted there are Messrs. Ogle, Dunn & Company, whose permission is herewith
acknowledged, and the Federal Reserve Board. See Chart XXXVI (3).
1074 BUSINESS CYCLES
(2) Brokers' Loans, New York City. From 1919 to 1926 the figures are for (call and
time) street loans placed by New York City reporting banks on their own account, and for
correspondents. For some banks these figures do not include loans to dealers in securities.
From 1926 to 1935 the figures are monthly averages of the loans to brokers and dealers on
security collateral, made by the weekly reporting member banks in New York City. They
represent call and time loans for own account, for account of out-of-town banks, and for
account of others, as compiled by the Federal Reserve Board. The two series, though very
imperfectly comparable, were spliced together. A much better one is available since
September 1935.
(3) Industrial Stock Prices. This index was taken, by permission^ from the Standard
Trade and Securities Service, Standard Statistical Bulletins. It is the monthly average
of their weekly stock price index (351 industrials). Weighting is according to the number
of shares outstanding. The average for 1926 = 100.
(4) Bond Yields. This index was taken by permission from the Standard Trade and
Securities Service, Standard Statistical Bulletins. It is an arithmetic average of yields to
maturity of 60 high-grade bonds, 15 industrials, 15 railroads, 15 public utilities, and 15
municipals. Monthly averages of Wednesday closing prices.
(5) Net Demand Deposits, New York City. See Chart L (1).
(6) Debits, New York City. See Chart L (3) [Outside Debits].
(7) Railroad Stock Prices. This index of 20 railroads, weighted by the number of
shares of stock outstanding, was taken by permission from the Standard Trade and
Securities Service, Standard Statistical Bulletins (e.g., see April 1934, p. 30). 1926 = 100.
(8) Loans and Discounts, New York City. See Chart L (5) [Outside Loans and
Discounts].
CHART LIII. ISSUES FOR "NEW CAPITAL," NEW YORK STOCK EXCHANGE.
ANNUAL
Figures are for total corporate, domestic and foreign, security issues for "New Capital"
by industrial groups, see Statistical Abstract of the United States, 1934, p. 278. Preferred
stocks of no par value and all common stocks are taken at their offering price, other issues
at par. See the Commercial and Financial Chronicle, Mar. 26, 1921, p. 1216, for a discussion
of the figures, of which it is the original source.
CHART LIV. UNITED STATES. LOG SCALE
(1) Failures. Liabilities, Monthly. Figures taken by permission from the Standard
Trade and Securities Service, Standard Statistical Bulletins, originally compiled by Messrs.
Ogle, Dunn & Company and others.
(2) Production of Producers' Goods. See Chart XLIII (1).
(3) "Productive" Capital Flotations. This series was used with the permission of
Moody's Investors' Service. Issues for refunding, acquisitions, working capital, and issues
of insurance companies, banks, etc., are excluded, as well as half of the issues for "Real
Estate Mortgages" and "General Corporate Purposes," which are classed as indeterminate.
They are in any case a very small percentage of the total. Federal issues are excluded,
but the figure for October 1934 includes 530 millions issued by the Home Owners* Loan
Corporation, because they were for the purpose of financing additions and improvements.
According to the terminology adopted in this book, the case for including this item is at
best a weak one.
(4) Corporate Earnings. The figures of earnings of all corporations, i.e., receipts less
statutory deductions, are from Statistics of Income, Report of the Commissioner of Internal
Revenue, Treasury Department, and have been corrected by W. L. Crum for dividends
received from other corporations. The author's thanks are due to Professor Crum for
APPENDIX 1075
kindly permitting the use of unpublished material. The items for 1931, 1932, and 1933,
being negative, could not be plotted.
(5) New Incorporations. Annual. Figures are for new enterprises with an authorized
capital of $100,000 or more. They were taken from the Statistical Abstract of the United
States, 1928, p. 309. After that year, this information was discontinued and would have
to be unearthed in the 48 statehouses, an undertaking much beyond the means at the dis-
posal of the writer.
CHART LV. GERMANY. LOG SCALE
(1) Call Rate. 1924-1935. This is the monthly average of the call money rate
(Taggeld), taken for 1925 to 1935 from the Institut ftlr Konjunkturforschung, Konjunktur-
Statistisches Handbuch, 1936, p. 112. The figures for 1924 are from the International
Abstract of Economic Statistics, 1919-1930, p. 93 (International Conference of Economic
Services).
(2) Stock Price Index. 1924-1935. This is the monthly index of the Statistische
Reichsamt (1924-1926 = 100). It covers about 325 stocks, including mining and heavy
industries, manufacturing, trade and transportation, and banking. See Otto Donner,
Die Kursbildung am Aktienmarkt, Vierteljahrshcfte zvr Konjunkturjorschung, Sonderheft
36, 1934. The Berlin Stock Exchange was closed from July 1931 to April 1932.
(3) Stock Exchange Turnover Tax Receipts. In order to get some sort of index of
stock exchange transactions, the monthly yield of the stock exchange turnover tax (taken
from the Institute fur Konjunkturforschung, Konjunktur-Statistischcs Handbuch, 1936,
p. 115) was adjusted for changes in the rate of taxation.
(4) Industrial Stock Issues (on the German Stock Exchanges). Annual figures taken
by permission from Die Wirtschaftskurve (a publication affiliated to the Frankfurter Zeitung),
vol. XIII, 1934, p. 267. The reference is to compilations of the Statistische Reichsamt.
For 1933 the figures include the Steuergutscheine, see Chap. XV, sec. F. The 1934 figure
is for the first 10 months.
(5) Industrial Bond Issues. See Series (4).
(6) Dividends. Per cent per annum. For source and method of calculation, see
Chart XXXVIII (5).
(7) Failures. Annual. 1925-1934. Figures are from the Statistisches Jahrbuchfiir
das Deutsche Reich, vol. LV, 1934, p. 368, and represent nominal capital of companies
instituting liquidation or bankruptcy proceedings (Einleitung dcs Liquidationsverfahrens
or Konkurserojfnung). Of course, this does not include reorganizations etc., which eco-
nomically may be equivalent to failures.
(8) Foundations. Annual. 1925-1935. Total share capital of new enterprises taken
from the Statistisches Jahrbuch fur das Deutsche Reich.
CHART LVI. GREAT BRITAIN. LOG SCALE
(1) London Town Clearings (as Distinguished from Metropolitan and Country Clear-
ings). Monthly figures for three weeks covering two stock exchange settlement days,
consols settlement day and the fourth of the following month. They were taken, by per-
mission, from the Monthly Bulletins of the London and Cambridge Economic Service.
(2) Industrial Stock Prices. The old index (1919-1924) is an arithmetic average of the
price relatives of 20 industrial ordinary shares, (1913 = 100). The last day of the month
to December 1922, and then the middle of the month. Weighting is based on 1907 Census
of Production. The new index (from 1924 on) is on the base 1924 = 100, and includes
prices of the shares of over 90 companies, operating principally or exclusively in the United
Kingdom (banks and railways are excluded). It is an arithmetic average of midmonthly
figures. Industrial groups are weighted according to the 1924 Census net output figures.
1076 BUSINESS CYCLES
individual stocks according to their market values in 1924. The new index was adjusted
to the old one by putting the 1924 average equal to 168 (see Special Memo 88 of the London
and Cambridge Economic Service, A New Index of Prices of Securities by A. L. Bowley,
G. L. Schwarz, and K. C. Smith, p. 14). Both indices were used with the permission of
the London and Cambridge Economic Service.
(3) London Call Rate. Day-to-day rate. Averages for the week ending the fifteenth
of each month, taken, by permission, from the London and Cambridge Economic Service
Monthly Bulletins.
(4) Money at Short Notice to Stock Exchange. Data are for 10 London Clearing Banks
from the Report of the Committee on Finance and Industry, 1981, p. 284. They do not
include money at call or money at short notice to the money market. * In the case of the
National Bank, Ltd., only figures relating to offices in England were included. Some items
in the earlier years are estimates.
CHART LVII. ISSUES FOR "NEW CAPITAL." GREAT BRITAIN. ANNUAL
New issues, domestic and foreign (conversions excluded), of stocks and bonds taken
from the London Economist, Annual Commercial History and Review.
(1) Railways. British, colonial, and foreign. In 1918 and 1919, only colonial and
foreign, and in 1916 and 1917 only British and foreign.
(2) Mines. Includes Australasian, South African, and others.
(8) Public Utilities. Include electric light, power, and telegraph; tramway and omni-
bus; gas and water.
(4) Shipping. Docks, harbor, and shipping.
(5) Iron and Steel. Iron, steel, coal, and engineering.
(6) Real Estate.
(7) Motors. Motor traction and motorcar manufacturing.
(8) Rubber.
(9) Oil.
CHART LVIII. UNITED STATES.
CONSPECTUS OF FEDERAL RESERVE BANK OPERATIONS.
ARITHMETIC SCALE
All the series were taken from the Annual Reports of the Federal Reserve Board and the
Federal Reserve Bulletins.
(1) Total Reserves. This includes nonreserve cash reported separately after Decem-
ber 1922. Total reserves are called due from the United States Treasury after January
1984. Gold included in these accounts is valued at $20.67 per fine ounce before January
1934 and $35 after that.
(2) United States Securities.
(3) Bills Bought.
(4) Float. From January 1919 to December 1926 this series represents the difference
between Uncollected Items and Deferred Availability, Notes of Other Reserve Banks being
included in Deferred Availability. From January 1927 to March 1936 the series repre-
sents the sum of Float as reported and Notes of Other Reserve Banks (separately reported
after December 1926).
(5) Notes in Circulation. The items included in this series are, from January 1919 to
May 1924, Total Notes in Circulation and Federal Reserve Bank Notes in Circulation;
from June 1924 to December 1926, Federal Reserve Notes in Circulation; from January
1927 to March 1936, Federal Reserve Notes Outside the Federal Reserve Banks, Federal
Reserve Notes Held by Other Federal Reserve Banks, and Federal Reserve Bank Notes in
Circulation.
APPENDIX 1077
(6) Bills Discounted.
(7) Government Deposits.
(8) Reserve Account of Member Banks Calculated — i.e.,
(1) + (2) + (3) + (4) - (5) + (6) - CO.
(9) Actual Reserve Account of Member Banks,
CHART LIX. UNITED STATES.
MAJOR RESERVE FACTORS AND CENTRAL MARKET RATES.
ARITHMETIC SCALE
(A) Total Reserves. See Chart LVIII (1).
(B) = A plus United States Securities. See Chart LVIII (2).
(C) = B plus Bills Bought. See Chart LVIII (3).
(D) - C plus Float. See Chart LVIII (4).
(E) = D minus Notes in Circulation. See Chart LVIII (5).
(F) = Bills Discounted. See Chart LVIII (6).
(G) Rate of the Federal Reserve Bank of New York. From Federal Reserve Bulletin.
(H) Commercial Paper Rate. Rate on 60- to 90-day prime commercial paper at New
York, taken, by permission, from the Standard Trade and Securities Service, Standard
Statistical Bulletins.
(I) Month-to-month Changes in Total Reserves.
(J) Yield on Government Certificates. Same source as Series (H).
CHART LX. BANK OF ENGLAND FIGURES. LOG SCALE
All figures are annual averages of Wednesday figures, as published in the Bankers'
Almanac and Yearbook and the Bank of England's Statistical Summaries. On Nov. 22,
1928, the note issues of the Bank of England and of the Currency Note Commissioners were
amalgamated, and the amount of the fiduciary note issue was fixed at 260 million pounds.
Therefore, figures of reserves and securities are not comparable. On Aug. 5, 1931, the
maximum of the fiduciary note issue was increased to 275 million pounds, and in April
1933 it was put back to 260 million pounds. No attempt has been made to present figures
about notes, reserves, or securities for 1928. From 1929 to 1934 Other Securities equal
Discounts and Advances plus Securities.
Index
AAA legislation, 988-992
Abbott, C. C., 813n.
Abel-Musgrave, R., 1037n.
Acceleration, principle of, 181-182, 916,
925
Accumulation, 75—84
definition of, 75
original, 229-230
(See also Corporate Accumulations)
Acquisitive principle, 495
Adaptation, waves of, 180, 183
Addison Act, 752
Advances of London clearing banks, 1931—
1937, 959
Agents de change, 679
Aggregate net income of corporations,
United States, postwar, 831-832
Aggregative theories of business cycles,
inadequacy of, 144
Agrarian depressions, in 1820's, 267-270
in last quarter of nineteenth century,
320-325
in 1920's, 732-743
Agricultural developments, 237-240, 266-
270, 319-325, 401-402, 732, 743
Agricultural machine industry, American,
317n., 389
Agricultural overproduction, 735n.
Akerman, Johan, 182
Alford, L. P., and Hannum, J. E., 802n.
Alfter, J., 95n.
All other accounts, 891
Allen, G. C., 374n.
Altschul, E., and Strauss, F., 738n.
American Telephone and Telegraph Com-
pany, index of general business, 23n.,
Amoroso, Luigi, 46w., 185
Anderson, B., 852n., 854«., 859w.
Anderson, O., 193n., 202n.
Andrew, A. P., 664n.
Angell, J., 598n., 886n.
Animal cycles, 530
Annalist, 26 n.
Anticipations, 53-55, 140
Antisaving attitude, 699-700
Applegate, LaRue, 1025
Armaments, 965
Armstrong, C. E., 165
Arthur, H. B., 459w.
Australian mining boom (see South African)
Austro-German customs union, plan of,
930-931
Autodeflation, 136
Automobile industry, rise of, 415-417
Autumnal drain, 656
Axe-Flinn index of business conditions for
Germany, 23 n.
Axe-Houghton index of business activity,
24n.
Ayres, Colonel, 12n., 166n., 490n.
Ayres, M. V., 826
B
Bagehot, W., 677n.
Baker, E. F., 785
Balance of payments, American postwar,
707-708
* This Index does not contain references to the numerous industries, concerns, and
individual entrepreneurs that have been mentioned in this book, especially in Chapters
VI, VII, XIV, and XV. Items that occur only in the Appendix have also been excluded*
1079
1080
BUSINESS CYCLES
Balance sheet of Federal Reserve banks,
combined, 888-894
Balances, demand for, 603
supply of, 606
Balfour Committee, 29 n.
Balogh, T., 680n.
Bank, assets, quality of, 116n.
clearings, 16
credit in national socialist Germany, 981—
982
debits, 16
epidemics, 912, 938, 943-944
failures, 16
Prussian, 308
rate, 16
in England, 1919-1929, 904-905
(See also Reserve bank rate)
rates in 1931, 933
reform of 1908, German, 400-401
reforms, nature of, 661
reserves, economy of, due to Federal
Reserve Act, 885n.
(See also Investments; Loans; Sec-
ondary reserves)
Bank of England, postwar policy of, 904-
905
rate of, 625-626
variations in gold stock of, 676-677
Bank of North America (1782), 293
Bank of United States, failure of, 911
Bank Act, of 1844, 307
of 1933, 987
of 1935, 1027
Emergency, 987
Bankers' acceptances, 624, 625
Bankers' banks (see Banks)
Bankers' loans (see Loans)
Bankhead Act, 991
Banking, commercial theory of, 115
investment theory of, 115, 862
reckless, 260
wildcat, 299-300
Banking series, postwar, 850-870
Banks, bankers', 112
central, 112
and crises, 660, 661
policy of, 648-665
and stock speculation, 689-691
Federal Reserve (sec Federal Reserve
Banks)
Banks, industrial or promoting in Germany,
348-350
land (see Landschaften)
member, 112
and Federal Reserve Board, 867-868
(See also Investments; Loans; Re-
serves)
mortgage, German (Hypothekenbanken),
359, 618
national, 310, 315, 316
special, in Germany, 406
of United States, 292
(See also Reichsbank)
Baring crisis, 381-382
Barnett, G. E., 418n.
Barns, A. F., 520n.
Barrett, D. C., 315n.
Barter terms of trade, 667
Baude, W. A., 215n.
Bauer, O., 696n.
Beach, W. E., 677n.
Beales, H. L., 353n.
Bean, L., 528, 532n., 823w.
Bellman, Sir H., 964n.
Below, G. v., 229n.
Belshaw, H., 321
Beneduce report, 718n.
Benner, S., 532n.
Bennett, M. K., 237n.
Bennett R., 56n.
Bercaw, L. O., 520n.
Berry, T. S., 458ra.
Bertrand, J., 60n.
Beveridge, Sir William, 166n., 179n., 458n.,
510n., 969n.
Bill brokers, English, 625n.
Bills drawn, Germany, 869
Bills of new tenor, 253n.
(See also Commercial; Finance; 'Treas-
ury)
Birck, L. V., 525n.
Bismarck's railroad policy, 346
Bituminous Coal Conservation Act, 933n.
Black, John D., 941n., 942n., 989n., 991n.,
992n.
Black Friday, 316, 336
Bland Act, 318
Bliss, C. A., 926n -928n.
Blodgett, R. H., 524n.
Bobroff's moving correlation, 202n.
Boehm-Bawerk, E. von, 78n., 223, 505,
548, 561, 836n.
INDEX
1081
Bond prices, Moody 's series of, 61 9n.
Bond yield, 16
behavior of, 618-622
index of (Standard Statistics Co.), 619n.
Bonds, collateral trust, 406
Borel, fimile, 194
Borrowing (see Government; Household)
Bortkiewicz, L. von, 452
Bouniatian, M., 220n.
Bowley, A. L., 59, 461n., 462n., 510n., 566,
729n., 751n., 799n., 847, 919, 967n.
Bowman, R. T., 830n.
Bradbury report, 723n.
Breaks in trend, 202
Bresciani-Turroni, C., 78n., 181n., 205,
877n., 982n.
Brokers (see Bill brokers; Stock brokers)
Brokers' loans, 16
Brown, E. C., 538n.
Brussels Conference, 824
Bryce, R., 456n.
Bubble Act, 248
Building, ahead of demand, 158n.
contracts awarded, 16, 19
permits, 16, 19
residential, 19, 159
societies, English, 753
Building boom, German (1872), 359
great English, 964-965
Building booms of 1920's, 743-753
Bullock, C. J., 181, 260n., 667n., 987n.
Billow duties, 401
Burgess, R., 888n.
Burns, A. F., 22n., 497n., 498n., 797n.
Burr, Ebersole, and Peterson, 831n.
Business, activity, 16
cycle theory (see Aggregative theories)
indices, 12n., 23, 24
normal, 4, 5
situations, 3
volume of, 16
Business Annals, lln., 221
Businessman's normal, 4
Buy-now campaigns, 587
Cairncross, F., 590n.
Call rate, 624
Call rates, England, 882-884
Germany, 880-881
New York, 876
Campbell-Bannerman, Sir H., 399, 727n.
Capacity (see Excess)
Capital, 42, 129
export, American, after World War, 703,
707
British, 367, 381
gains tax, 1040
goods (see Interest)
import German, postwar, 718-722
issues (see Issues)
market, 129
Capitalism, competitive, 96
definition of, 223
rise of, 228
stationary, 1033
trustified, 96, 145
Capitalist and entrepreneur, 103-104
Caprivi treaties, 308
Carli, G. R., 451
Cartel quota, 62
Cash held by corporations, 860n.
Cassel, Gustaf, 38, 472, 680n.
Causal factors, 25
Central banks (see Banks)
Central creation (see Credit)
Central market, 126
American, theory of, 892-894
Certified check accounts, 680
Chain method, 456
Chain stores, sales of, 16, 19
Chamberlain Act, 752, 918
Chamberlin, E. H., 56n., 63, 65n., 67n.
Champernowne, D. G., 80n., 968n.
Change, autonomous, 14
conditioning of, 72
external factors of, 6-13
internal factors of, 72-87
Changes, in consumers' tastes, 73-74
structural, lln., 12n.
Changing harmonics, 18 In.
Chapman, J. M., 650n.
Chawner, L. J., 476n., 745n.
Chinese indemnity to Japan, 677
Circulation, velocity of, 545-546
(See also Currency)
Civil War, 334
deflation after, 315-317
economic effects of, 314-315
Clapham, J. H., 252n., 254, 274n., 276,
303n., 367n., 565n.
Clark, Colin, 729n., 765n., 798, 799, 828-
829, 849n., 850, 885, 97 On.
1082
BUSINESS CYCLES
Clark, G. N., 233n.
Clark, J. B., 64
Clark, J. M., 560n., 806n.
Clark, V. S., 231n., 252n., 290, 384, 409n.,
415n., 421n., 565
Class consciousness, 699n.
Clausing, G., 859n., 766
Clean cyclical series, 200
Clean trend series, 200
Clearing-house certificates, 334w.
Cleveland Trust Company, business index
of, 12n., 23n.
Coal famine, 413
Coal mining, 242
Coats, R. H., 460n.
Cobb and Douglas, 590rc.
Cobweb problem, 49
Coffee cycles, 530
Cole, A. H., 252n., 392n., 471, 562n., 582w.
Collateral loans, 18, 583
Colm and Lehmann, 1002n., 1039n.
Colonial Stock Act (1900), 430n.
Columbia Committee on Economic Recon-
struction, 802n.
Commercial and Financial Chronicle, 26 n.
Commercial bills, 15
decreasing importance of, 625
Commercial crisis, 257
Commodity credit corporation, 991
Commodity stocks, 524
Communist Manifesto, 497
Companies Registration, Act of 1844, 307
Competition, cutthroat, 61
imperfect, 56-68
monopolistic, 62, 63-68
perfect, 46
wastes of, 60
Condliffe, J. B., 924n.
Conklin, W. D., 476n.
Consequential factors (series), 25
Consumers' credit (see Public spending)
Consumers' expenditure, 236, 299
and realized income, postwar, United
States, 823-827
Consumers' prosperity, 718, 721, 811
Consumption, 15
Continental paper currency, 253n.
Continental system, 257
Copeland, M. T., 393n., 819, 820n.
Corporate accumulations, United Kingdom,
postwar, 829
United States, postwar, 820-822
Corporate state, 697
Corporate surplus and real investment, 822
Corporation tax (see Tax)
Corsten, H., 222n.
Cost, decreasing, law of, 90
increasing, law of, 90
of living (see Price level)
Cost curves, 89-91
Costs, index of, 476
Cotton mania, 288 %
Cotton program, effects of, 991-992
Cournot, A., 59, 60n.
Cox, G. V., 22n.
Crandall, R., 458n., 470n.
Credit, absorption of, by stock speculation,
680-681
central creation, 650
autonomous, 893
responsive, 893
member creation, 650
rationing of, 651
(See also Bank credit; Commodity
credit corporation)
Credit creation, meaning of, 110-114
and railroads, 329
Credit mobilier, 347n.
Crises, 162
English, in seventeentn and eighteenth
centuries, 249-250, 296, 297
no technical meaning of, 5
theories of, 162
(See also Baring, Commercial, and
World crises; Central banks; Waren-
handels Krisen)
Crisis of concerns, in Germany, 763
"Crisis of the dollar," 937
Crisis of 1810-1811, 262
of 1815, 262, 298
of 1818, American, 301
of 1825, 279, 299
of 1836-1839, 299-300
of 1847, in England, 344-345
in Germany, 350
of 1857, 331
in England, 377
in Germany, 361-362
of 1866, in England, 378
of 1873, 336-337
in England, 380
in Germany, 362-364
of 1883-1884, 340
INDEX
1083
Crisis of 189S, 341
in America, 388-389, 402, 403
in England, 382
in Germany, 366
of 1896, 317
of 1907, in Germany, 448
in United States, 407, 408, 410, 424-
429
of 1921, in United States, 786
Crops, variations of, 8, 176-179
Croston, J. J., and Davenport, D. H., 841n.
Crum, W. L., 20, 165, 166n., 169n., 181,
210rz., 212w., 498w., 557w., 615n., 629,
814, 819, 830n., 832, 833/t., 924n.,
1006n.
Crum and Patton, 193n.
Cumulation, 181
Cunliffe report, 723
Cunningham, W., 254n.
Currencies, gold, postwar, 703
Currency, in circulation and deposits,
ratio of, 886-887
Currency school, 640
Current accounts, postwar behavior of, 870
Currie, L., 851n.-853n., 855n., 856«.,
859n., 900n., 996n., lOOlrc., 1002«.,
1012n.
Cycle, formal definition of, 200
Cycles, 22, 138
multiplicity of, 161
(See also Animal; Coffee; Four-phase;
Harvest; Hog; Juglar; Kitchin;
Kondratieff; Special; Specific; and
Shipbuilding cycles
Cyclical industries, 158
Cyriaci-Wantrup, S. von, 618
D
Dailey, D. M., 854n.
Darmois, G., 193n.
d'Avenel, Vicomte, 458w.
Davenport, D. H., and Croston, J. J., 841n.
Davis, J. S., 989n., 991n.
Dawes loan, 715
Dawes plan, 702-704
Day, E. E., 490«.
Debt-deflation, 146, 909, 925
Debits, 1014
and deposits, postwar relation of, 857-858
and loans, postwar relation of, 858-859
postwar behavior of, 817-820
Deferred availability, 890n.
Deflating of monetary expressions, 456
Deflation, 260-266
after Civil War, 315-317
in England, postwar, 725
in Germany, in 1930, 923
in 1931, 935
Deflation and inflation, 260
Deflationary effect of saving, 587
Deflationary effects of social security
payments, 1032n.
Deflationary pressure, 939-940
Deflection, 579
Demand, for durable goods (see Interest)
for money, 548
replacement of, 158, 189-191
Demand deposits (see Deposits)
Demonetization of silver, in Germany,
308-309
Department stores, sales of, 16
Deposits, 16, 950, 1007
compensated, 120
created, 113
and currency (see Currency)
and debits (see Debits)
demand, adjusted, 851
net, 851
and loans (see Loans)
in London clearing banks, 1931-1937, 959
original, 119
outside, payroll covariation with, 573 n.
owned, 119
time, 583
nature of, 850-857
velocity of, 579, 585, 856
Depression, 149
deep, 154
Great, 1873-1896, 353-354
(See also Agricultural and Agrarian
depressions)
Depressionless industries, 1010
Descriptive trends, 201
Determinateness of economic quantities, 41
Devaluation, American, 998
and stock prices, 1000
Devons, E., 799n.
Dewey, D. R., 315w.
Dieterici, K. F. W., 252n.
Discount houses, 625n.
Discounting, new system of, 625
process of, 609
of quasi-rent, 61 In.
1084
BUSINESS CYCLES
Discounts, 16
and loans (see Loans)
Discoveries, 8-9
Disequilibrium, 47
Disorganized markets, 540
Dispersion graphs, 521 n.
Disproportionalities in price movements,
948-950
Dissaving and gains from stock trans-
actions, 679
Dividends, 17
and earnings in 1931, 951-952
Divisia, Francois, 452
Dodge Corporation, 745n.
Donner, Otto, 21 n.
Douglas, P. H., 510n., 565, 798, 799n., 88671.
Douglas and Cobb, 590n.
Dumping, 61
Dunbar, Charles F., 220n., 830, S34n.
Dunlop, J., 277n.
Duopoly, 60-61
Dutot, 451
Dyer and Martin, 894n.
Dynamics, 48
E
Earnings and dividends (see Dividends)
Easy money, 17
(See also Stock pricing)
Easy money policy, 658
Ebersole, Burr, and Peterson, 831n.
Economic evolution, 86
Economic horizon, 40, 99
Economic policies in America, 1981-1932,
939-943
Economic system, 41
Economist's Index of Business Activity,
24n.
Econostat, business index of, 23n., 24n.
Eddy, G. A., 879n.
Edgeworth, F. Y., 50, 60n., 62, 452, 459n.
Edie and Weaver, 585n.
Edison Electric Institute, 411
Efficiency of money, 546
Effort elasticity, 159n., 177, 324
Ehrenberg, V., 225n.
Ehrke, K., 476n.
Eiteman, W. J., 872n.
Elastic waves, 179-183
Elasticity, product, 38n.
of substitution, 80
Electricity Supply Act, of 1919, 43S
of 1926, 758
Emergency Banking Act, 987
Emergency Farm Relief Act, 1933, 987, 990
Emergency Relief Act, 1933, 1001
Employment, 16, 509-519, 1009, 1011, 1015
Germany and United States, 1931 and
1932, 945-946
in national socialist Germany, 971
Enclosures, 238
Endogenous factors, 7n.
theories, 7n., 48
England, 1931-1938, 954-971
Entrepreneur, enterprise, 102-104
Entrepreneur and capitalist, 103-104
Epstein, R. C., 415n., 416, 773n., 830n.,
832»., 833n., 834
Equilibrium, aggregative, 48
concept of, 80-71
general or Walrasian, 42
imperfect, 44
neighborhood of, 71, 173
partial or Marshallian, 43
perfect, 44
sloppy, 44
state of 42
use of, 68-71
zone of, 58
Equilibrium values, 42
Ernie, Lord, 267n.
Erratic series, 19
Errors of judgment, role of, 140
Essars, P. Des, 586
Estate tax (see Tax)
Evans, G. C., 57n., 184n., 528
Evolution, economic, 86
Excess capacity, 54, 61, 62, 66, 67, 509, 754,
802-808
reserves (see Reserves)
Exchange, equation of, 44
foreign (see Foreign exchange)
Exchange equalization account, 957n.
Exogenous factors, 7n.
Expectations, 53-55, 140
Expenditure of households, relation to
household's receipts, 549n., 550n.
and relief in 1931-1932, 942-945
(See also Consumers', Federal, System
expenditures)
Exploitation, theory of, 229
Exports, 15
British, 367
INDEX
1085
Exports, British, postwar, 755-756
postwar, from England to Germany, 703
External economies, 92, 239, 875, 390
External factors, 6-13, 700
Ezekiel, M., 520n,, 522n., 532n., 736n.,
819n., 951n.
Fabricant, S., 821, 831n., 833, 913n., 950n.
Failures, 16
Farm Credit Act, 988n.
Farm Mortgage Corporation Act, 987
Farming situation in 1931 and 1932, 941-
942
Fascist state, 697
Feavearyear, A. E., 829
Federal expenditure, postwar, 709
Federal reserve banks, combined balance
sheet of, 888-894
Federal Reserve Board, and member banks,
867-868
Federal Reserve System, assistance ren-
dered to European Central Banks,
888n.
mechanics of, 888-894
policy of, 894-904
and causation of world crisis, 902
Fetter, F. A., 497
Finance bills, 624
Financing, primitive, 230
Finanzkapitalismus, 405n.
Fiscal policy, in bourgeois era, 310-311
postwar, 709-714, 716-722, 729-731
in United States since 1932, 1038-1042
Fisher, Irving, 35n., 127n., 129n., 143n.,
146-147, 185, 212n., 452, 455n., 457,
459n., 470, 490n., 497, 509n., 522n.,
584n., 746w., 900n., 909w., 938
Fisher, R. A., 193n., 202n.
Fisher's ideal formula, 456
Flight from the land, 325
Float, 890
Flux, A. W., 799n.
Foenus nauticum, 615
Folio w-the-leader system. 61
Forced savings, 112n.
Fordney-McCumber Act, 706
Forecasting, 13
Foreign exchange, 17
Foreign exchanges and differences in money
rates, 674
Foreign invesments (see Investments)
Foresight, 52
Forethought, 52
Forty-months cycle, 165
Foster, R. R., 745n.
Foundations of firms, 17, 94
Fourier analysis, applicability of, to time
series, 198-199
Four-phase cycles, 149
Franklin, B., 287
Fraser, L. M., 877n.
Free gold, 887-888
trade, in England, 306-307
Freehand methods of analysis, 21 In.
Freight-car loadings, 17
French, D. R., 854n.
French indemnity, 1871, 313-314
Frickey, Edwin, 22n., 23n., 200n., 202n.t
203, 204, 316n., 551n., 615n., 877n.
Friction, 42n., 50
Frisch, Ragnar, 48, 60n., 151n., 171n., 181.
184n., 189, 195, 208-212, 215-219, 452,
469, 560n.
Fullarton, 640
Gay, E., 224n., 238
Gayer, A. D., 916n., 942
Gehlhoff, W., 462n.
Georgescu-Roegen, N. S., 215
Gibson paradox, 633n.
Giffen, R., 590n.
Gilboy, E. W., 469n., 527n., 529n., 564n.,
565
Gilfillan, S. C., 85n., 227n.
Gill, C., 745n., 749n.
Gladstonian finance, spirit of, 311
Glass-Steagall Act, 939-940
Goether, E. T., 538n.
Gold, free, 887-888
Gold account, inactive, 1028
Gold-buying policy, 998
Gold currencies, postwar, 703
Gold devices, 676
discoveries, 8
flows, 17
influx into United States, 863n, 887-888,
997, 1028
market, London, 675-676
and price level, 472-473
production, variations in, 176
1086
BUSINESS CYCLES
Gold Reserve Act, 997-998
scare, 667n.
South African, 676
Standard Act, 317
of 1925, 726
Gold standard, England's abandonment of,
955-958
postwar working of, 887-888
Gold Standard Suspension Act, 955
Gold sterilization, 888n., 1027
Golden weiser, E. A., 71 In.
Goodrich, C., 805n.
Gordon, R., 888n.
Goschen, Lord, 382, 607, 675
Government borrowing, cyclical behavior
of, 607
Government spending in national socialist
Germany, 975-977
Gray, T. H., and Terborgh, G. W., 746n.
Greenstein, B., 166n., 171n.
Gresham, Sir Thomas, 234n.
Growth, 83-84, 158
(See also Organic growth)
Gruenbaum, F., 550w.
Guyot, Yves, 15
H
Haberler, Gottfried von, 452, 454n.
Hahn, A., 582n.
Hakluyt, R., 246
Haldane, J. B. S., 184n.
Hall, L. W., 583n.
Hamilton, Alexander, 287, 288
Hamilton, E. J., 225n., 231, 252n., 459n.,
564n.
Hammond, Bray, 260n., 294n.
Hanau, A., 532
Hanch, C. C., 826n.
Hannum, J. E., and Alford, L. P., 802n.
Hansen, A. H., 78w., 5COn., 565, 634n.,
907?i., 1033n.
Hardy, C. O., 601 n., SSSn.
Hardy, C. O., and Owens, R. N., 680w.
Harmonics, changing, 181 n.
Harms, Bernhard, lln.
Harris, S. E., 758«., 866n., 888n., 958n.,
961n., 996n.
Harrod, R. F., 43n., 60n., 64n., 67n., 68,
144, 954w.
Harvard method, 21-24
Harvest cycle, 176-179
Hatry failure, 908
Hawley-Smoot Act, 707
Hawley-Smoot tariff, 915
Hawtrey, R., 142, 634n., 724n., 904n., 918n.,
956n.
Hayek, F. A. von, 78n., 296n., 333, 345n.,
603n., 634
Hayek effect, 812, 814
Heckscher, E. F., 245n.
Heichelheim, F., 226n. %
Hepburn, A. B., 583n.
Hesitations, 183, 428
Hicks, J. R., 56n., 60n., 65n., 80n., 574n.
High-pressure engine, 289
Hilferding, R., 405n., 696n., 715n.
Hirsch, J., 760
Historical variables, 194-197
Hoarding, 76, 579
Hobson, C. K., 367n., 429
Hoffmann, W., 490n.
Hog cycle, 49, 54, 531-533
Holmes, C. L., 739n.
Home Owners' Loan Acts, 987
Hooker, L., 572n.
Hoover, E. M., 391n., 786n., 788n.
Horizon, economic, 40, 99
Hoskins, L., 244n.
Hotelling, H., 63n.
Households' borrowing, postwar, 826-827
Housing Act of 1936, 1025
Hubbard, J. B., 945n.
Hudson, railway king, 344
Hunt, B. C., 278n., 279n.
Hutchinson, W. T., 389n.
Hyndman, H. M., 162n.
Hypothekenbanken (see Banks)
Hytten, J., 147n.
Illinois Central Railroad, 330-331
Imperialism, 432
Neo-Marxist theory of, 405n.
sociology of, 696n.
Imports, 15
Inactive gold account, 1028
Inactive systems, 42n., 160
Income, aggregate net, of corporations,
United States, postwar, 831-832
generation of, by public expenditure,
920n., 923, 1001-1006
national, 561, 950, 1001-1002, 1014
INDEX
1087
Income, parities, 736
postwar behavior of, in Germany, 827
in United Kingdom, 828
in United States, 819
realized (see Consumers' expenditure)
tax (see Tax)
velocity of money, 545
Incomes, sum total of money, 15
Independent Treasury Episode, 293n.
Indeterminateness, 50
India, native capitalism in, 705
Indices of business conditions, American
Telephone and Telegraph Company,
23n., 24n.
Axe-Flinn, 23n.
Axe-Houghton, 24n.
Cleveland Trust Company, 12n., 23«.
Economist, The, 24n.
Econostat, 23n., 24n.
New York Times Weekly, 23n.
Ogburn-Thomas, 24w.
Individual series, 25
Industrial Advances Act, 987n.
Industrial policies in United States after
1933, effects of, 1043-1044
Industrial revolution, 168, 170, 253
New, 397
of 1920's, 753-794
Inflation, and deflation, 260
safeguards against, 997n.
Inflation Act, 997-998
Inflationist interests and government, 986w.
Innovation, 84-86
resistance to, examples of, 243, 244
theory of, 87-192
Instalment sales, 826
Interest, 27, 32, 41, 123-126
adapted rate of, 603, 605, 636
and demand for durable goods, 609
equilibrating rate of, 603w., 605
equilibrium rate of, 603
lag in, 605
natural rate of, 126
postwar behavior of, 811-817
and quasi-rent, 608
rate of, 602-638
real rate of, 126
and value of capital goods, 609
Interest rates, and issues, 880
on mortgages, 617-618
structure of, 611
Intermediate products, 38
Internal economies, 92
Internal factors, 7n.
International relations, cyclical aspects of,
666-678
Inventions, 8-9, 84-86
Investment, 76
in England, postwar, 885
real, 76
and corporate surplus, 822
issues for, 879-880
temporary, 75n., 77, 580, 606, 623, 644,
870-872
Investment opportunity, theory of vanish-
ing, 1032-1036
Investment survey, Moody's, 880
Investment theory of banking, 115, 862
Investments, foreign, English, behavior in
world crisis, 962-963
and loans, of banks outside New York
City, 645-646
member banks', 595, 598, 862-868, 870
Irregularities, external, 7n., 144
internal, 144
Issues, capital, English postwar, 883-885
and interest rates, 880
for New Capital, United States, postwar,
877-879
for real investment, 879-880
of securities, 15
in national socialistic Germany, 983
Japan, industrialization of, 705
Jenks, J., 367n.
Jerome, H., 402, 419, 747n., 783n.
Jevons, W. St., 176, 371, 452, 461n., 469n.,
627n., 656w.
Joint Stock Companies Act, 307
Jones, G. T., 476w.
Juglar, Clement, 139, 162, 163, 326n.
Juglar cycles (Juglars), 169
K
Kahn, R. F., 80n., 92, 515, 752n., 758
Kaldor, N., 50n.
Kalecki, M., 185-188
Karsten, K. G., 181, 877n.
Kemmerer, E. W. A., 490n., 584
Kerr-Smith Act, 990
Keyes, E. W., 583n.
1088
BUSINESS CYCLES
Keynes, J. M., 17, 43, 127n., 585n., 608n.,
633n., 832n., 838n., 951».
Kindersley, Sir R. M., 962n.
King, Gregory, 237
King, W. J., 490n., 562n., 820n., 823
King, W. T. C., 292, 378n., 626n.
Kitchin, Joseph, 165, 169n., 215, 473, 555n.
Kitchin cycles (Kitchins), 169
Knickerbocker failure, 428
Knight, F. H., 52
Kondratieff, N. D., 164, 470n., 477n.
Kondratieff cycles (Kondratieffs), 169
Konsum-Konjunktur, 718
Kreditanstalt, Austrian, breakdown of,
980n.
Kreps, Thomas J., 809n.
Kreuger failure, 908
Kuczynski, R., 565ra.
Kuznets, Simon, 21n., 165, 389, 398n.,
497n., 500n., 520n., 801, 819, 823, 912,
950n., 951, lOOln., 1002n.
Labor policies in United States after 1933,
effects of, 104&-1043
Labor Relations Act, national, 1042, 1045
Lag, of maximum correlation, 502
and sequence, 57n.
Lag distribution, method of, 509re.
Lags, technological, 48
Land grants to railroads, 328
Land sales, receipts from, 326
Land settlement, policy of, 319
Landschaften (landbanks), 618
Prussian, 268n.
Lange, Oskar, 49n., 520n.
Laspeyres' formula, 456, 490
Lausanne agreement, 936
Lavington, F., 54n., 430n.
Law, John, 250-252
Layton, Sir W., 461n.
Layton report, 718n., 721
LeChatelier, principle of, 47
LeCorbeiller, C., 210n.
Lederer, W., 827n.
Lehmann and Colm, 1002ra., 1039n.
Leonard, R., 237n.
Lenoir, M., 501n., 520n.
Leonard, £., 490n.
Leontief, W. W., 49n., 60n., 452, 476n.,
520n., 521n., 575n., 822
Leong, Y., 490n., 584n., 796n.
Lerner, A. P., 63w.
Less-than-carload-lots, 17
Levin, Moulton, and Warburton, 820n.
Lidderdale, W., 626n.
Limited Liability Act, 307
Lippmann, W., 916n.
Liquidation, abnormal, 149
Liquidations, 15
List, Friedrich, 115
Loan, Dawes, 715
Loans, on account of others, 872-874
all other, postwar behavior of, 859-860,
912
bank, 16
bankers', boom, 791
postwar theory of, 872-874
brokers', 16
collateral, 18, 583
and debits, postwar relation of, 858-859
and deposits, covariation of, 594-597
real estate, member banks, 855, 859
total outside, and discounts, 802
London Economist, 26 n.
London gold market, 675-676
Long wave, 164, 168
Lorenz, P., 201n.
Ltfsch, August, lOrc., 75n.
Lotka, A. J., 492n.
Lough, W. H., 821n., 823, 824n., 825n., 826,
851n.
Lowe, Adolf, lln.
Lowell, F. C., 288n., 290
Lubin, J., 806n.
Lumpiness of factors, 39, 80, 92
Lyon, L. S., and associates, 994n.
M
Macaulay, F. R., 201n., 619n., 813n.
McCulloch, Secretary Hugh, 315
McGroarty bill, 71 In.
Machlup, F., 136n., 333n., 634n., 680n.
McKenna duties, 706
McLeish, A., 806n.
Macrodynamics, 185
Mail-order houses, 15
Mantoux, Paul, 253n., 297
Manufactures (see Output)
Marginal degree of productivity, 38n.
Marriage rates, 10, 74
Marschak, J., 760
INDEX
1089
Marshall, Alfred, 36, 43, 45, 92, 148, 456,
714n.
Marshall-Moore theory of organic growth,
203
Martin, R. F., 802n.
Marx, Karl, 7, 10, 104, 189, 190n., 223, 229,
497, 695, 698
Mason, E. S., 395, 988n., 994n., 1020w.
Mata, Garcia, 8n.
Matthews, A. M., 686n.
Maverick, L. A., 21 In.
Median-link-relative method, 20n.
Member banks (see Banks)
Member creation (see Credit)
Mercantilism, 233-235
Merchandise transported, miscellaneous, 17
Merchant adventurers, 246
Merger movement, 403-411
Merton, R. K., 85n.
Migrations, 10
Mill, J. S., 78n., 314n.
Mills, F. C., 51n., 193n., 201, 453n., 459n.,
476, 504, 520n., 522n., 526n., 527,
529, 532, 745?i., 796, 797n., 801, 804,
805n., 809n., 814n., 830rc., 841n., 859n.,
926n., 928n., 947n., 948n.
Miquel, Johannes von, 311
Mises, L. von, 634n.
Mises, R. von, 193n.
Mitchell, W. C., 22n., 33n., 39n., 139, 144n.,
162, 163n., 164n., 165, 178n., 202n.,
221n., 224, 315n., 459n., 476n., 480w.,
482, 498n., 568, 583n., 584, 666n.,
790n., 803n., 830n., 840
Monetary expansion, in England, 958-959
Monetary expressions, deflating of, 456
Monetary ligamen, 44
Monetary management, in national socialist
Germany, 979-981
in United States, 1932-1935, 996-1001
Monetary metals, influence of supply of,
231-233
Monetary monomania, 233 n.
Monetary parameter, 453
Monetary strain in 1928-1929, 876rc.
Money, in circulation, 583
efficiency of, 546
outside of banks, 16
stamped, 587
tight, 17
velocity of, 545-546
Money market, 126, 612
tension in, 17
Money rates, 16
differences in, and foreign exchange, 674
in and after crisis, 944-945, 999, 1008,
1016-1017
Money Trust, 639
Monopolies, 244, 245?t., 246n.
Monopoly, 40, 56-57
bilateral, 57-59
universal, 57
Monopsony, 57
Moody's Investment Survey, 880
Moody 's series of bond prices, 619n.
Moore, H. L., 38n., 45, 49w., 69, 166n., 176,
178n., 526, 527, 609
Moratorium, Hoover, 702, 933
Morgenstern, O., 53n.
Mortgage banks (see Banks)
Motorcar sales, 17
Moulton, H. G., 601n., 820rc.
Moving correlation, Bobroff s, 202n.
Mund, V. A., 541n.
Myrdal, Gunnar, 460n., 603n., 838n.
N
National banks (see Banks)
National grid, English, 758
National income (see Income)
National Labor Relations Act, 1042, 1045
Natural series, 18
Navigation Act of 1489, 234
Necco index of prices, 462n.
Negotiability, importance of, 613
Neisser, H., 674n., 675, 817n.
Neo-Marxian theory, 432
Neomercantilism, 898, 696
English, 960-961
Nerlove, S. H., 830n.
Net income velocity of money, 545
New capital issues (see Issues)
New countries, discovery of, 8-9
New Deal policy, 986-1050
New eras, 173
New middle class, 698
New security listings, 599
New York Times Weekly Index, 23n.
New Zealand, postwar conditions in,
704-705
Newmarch and Tooke, 343n., 461n,
Nichol, A. J., 60n.
1090
BUSINESS CYCLES
Norfolk system, £66
Norm, theoretical, 45, 70
Normal, business, 4, 5
businessman's, 4
statistical, 206
Normal business situations, 4, 5
Normal points, method of, 208-210, 215-
219
Normal values, 45
Norris-LaGuardia Anti-Injunction Act, 994
Northern Pacific Corner, 407, 426
Norton, J. P., 582ft.
Nourse, E. G., 777n., 989n.
Noyes, A. D., 582n.
NRA legislation, 992-996
NRA policy, 539
NRA wage policy, 994
Occupational statistics, German, 854n.
Ogburn-Thomas Index of Business Cycles,
24ft.
Ohio Life and Trust Company, failure of,
332
Ohlin, B., 924ft.
Old-age reserve account, 1042
Oligopoly, 60-61
Olivier, M., 520n.
Open market, 126
Open market operations, 916-917
history and effects of, 896-904
Open market rates, 622-625
Optimism, 141
Organic growth, Marshall-Moore theory of,
203
Orton, W., 430?i., 432
Oscillations, 180, 183
Ottawa agreements, 962-963
Output, 1009-1011, 1013-1014
English, 1931-1938, 966-967
German, 1933-1938, 971-972
of manufactures, behavior of, 1919-1929,
795-799
total, concept and measurement of, 484-
491
prewar behavior of, 491-509
in world crisis, 924-930
Outside clearings, analysis of, 548-561
Overdrafts, 583
Overend-Gurney and Co., 378n.
Overinvestment theory, 141
Overproduction, 141
agricultural, 735n.
of equipment in 1920* s, 801-802
Oversaving theories, 820ft., 822, 826-827
Overspending, 579, 585
Overstone, Lord, 640
Owens, R. N., and Hardy, C. O., 680n.
Paasche's formula, 456
Paetzmann, H., 742n.
Paish, G., 563w.
Palander, T., 60n., 63n.
Palgrave, R. J., 626ft.
Panic, rich men's, 407, 426
Pantaleoni, Maffeo, 48
Papen plan, the, 935-936
Pareto, Vilfredo, 47, 49ft., 52
Paton, W. A., 830n.
Payroll, 1009, 1015
covariation with outside deposits, 573n.
tax, 1042
(See also Wage bill)
Payrolls, postwar behavior of, Germany,
827
United Kingdom, 829
United States, 819-820
Pearson, F. A., 323
Pearsons, C. E., 746ft.
Peel's Act, suspensions of, 661
People's budget, 727n.
Pereire, brothers, 347ft.
Periodicity, 143
Periodogram analysis, 165, 166n., 169
Periodograph, 165n.
Perpetuum mobile,, 139
Persons, Warren M., 23n-25n., 181, 205,
316ft., 411, 459n., 475n., 490, 491n.,
509ft., 510, 525, 596n., 621, 622n.,
786ft., 877ft., 907ft.
Pervushin, S. A., 178ft.
Pessimism, 141
Peterson, Ebersole, and Burr, 83 In.
Pfandbriefe, 268ft., 618
Phinney, J. T., 473n.
Pig-iron consumption, 485-487
Pigou, A. C., 56, 63, 80n., 92, 123, 177n.,
452, 509ft., 510, 519n., 585, 850n., 954ft.
Plenge, A., 347n.
Political Arithmetick, 15w.
Polk, R. L. and Co., 801n,
INDEX
1091
Pollard, Spencer, 667n.
Population, changes in, 10, 74
Porter's Progress of the Nation, 252n.
Posner, H. L., 806n.
Potosi mines, 231
Poverty in plenty, 365
Pribram, Karl, 71n., 539n.
Price, maintenance (see Resale)
movements, disproportionalities in, 948-
950
parities, 736n.
system, 453
Price level, 16, 449-482
and cost of living in England, 1931-1938,
969-970
and gold, 472-473
theory of, 452-458
world's, 462n.
Price levels, postwar behavior of, 807-811
Prices, 16
behavior of, in 1932 and after, 947-950,
1008, 1018-1020
delivered, 538
group, 475-476 *
in national socialistic Germany, 977-978
rigidity of, 57, 536
secular fall in, 465
sensitive, index of, 23n.
stickiness of, 57
Primary factors, 25
Primary series, 25
Primitive financing, 230
Privat-Diskont, 625
Product elasticity, 38n.
Production, 15
coefficient of, 38n.
cost of, 15
function of, 38, 39
method of, 39
Production process, synchronization of, 40
Productive combination, 38
Productivity, 38n.
marginal degree of, 38n.
Profit, declining rate of, law of, 628
Profitless prosperities, 105
Profits, 15, 27, 105-109, 1006, 1014-1015
postwar behavior of, 830-834
Prognosis, 13
Progress, 86, 102
Promoters' epoch, German, 362
Propagations, 189, 219n.
Prosperity, 138
See also Consumers' prosperity; Prof-
itless prosperities)
Prosperity plateau, 135
Protection, influence of, 258-260
Protectionism, postwar, 705-708
Public spending, and consumers' credit,
1004
and recovery, United States, 1001-1006
and slump, United States, 1031-1032
Public Utilities Holding Company Act, 1043
Puddling process, 273
Pulse charts, 464
Pure model, 138
Q
Quantity theory, 44
Quasi-rents, 41
discounting of, 61 In.
and interest, 608
Quittner-Bertolasi, Ellen, 200n.
R
Railroad amalgamations, early English, 344
Railroad and Canal Traffic Act of 1888
(England), 342
Railroad earnings, 17
Railroadization, 168, 325-351
Railroads, 158, 402-403
and credit creation, 329
land grants to, 328
Railway mania in England, 343
Random or stochastic variables, 194-197
Rationalization, index of, 88n.
postwar, in Germany, 759
Rayon industry, rise of, 433—435
Real estate loans (see Loans)
Real estate market, 16
Real investment (see Investment)
Real trend, 204
Recession, 138
of 1927, 790-791
Reciprocal trade agreements, 1026
Reconstruction Finance Corporation, 940
Recontracting, 50
Recovery, 149
natural and sound, 995-996
(See also Public spending; Reform)
Recovery point, 190
theory of, 151-154
1092
BUSINESS CYCLES
Recovery policy, in America, 983-1011
in nationalsocialistic Germany, 974-975
Recurrence of great wars, 697n.
Rediscounting for brokers, in England, 626
Redlich, Fritz, 222n.
Reference trend, 204-205
Reform versus recovery, 985n.
Reichsbank, 308
in 1931, 933-934
Reichskaasenscheine, 678
Reisch, R., 680n.
Reithinger, A., 803n.
Relief expenditure in 1931-1932, 94fc-945
Renner, K., 696n.
Reparation problem, discussion on, 704n.
Replacement, 158, 189-191
Resale price maintenance, 538
Reserve account, member banks, 891
Reserve bank rate, theory of, 895-896
Reserve ratios, 16
Reserves, excess, 939, 999,1028-1031
of member banks during 1920's, 868
secondary, of banks, 644
Response, 72
apparatus of, 68, 72
Result trend, 206
Resumption Act, 316
Retail trade, 16
Retardation, 497-500
Revival, 149
Rhodes, Cecil, 381
Rhodes, E. G., 967n.
Ricardo, David, 88
Ricci, Umberto, 49n.
Richards, H. J., 989n.
Richards, R. D., 226n., 249n., 292
Riefler, W., 888n.
Riggleman, J. R., 744n.
Risk bearing and enterprise, 104
Robbins, L., 924n.
Roberts, J. B., 95n.
Robertson, D. H., 39, 78n., 91n., 112n.,
159n., 356
Robinson, Joan, 43n., 56n., 57n,, 63, 80n.
Rodbertus, K., 325n.
Rogers, H., 681n.
Rogers, Thorold, 458n., 564n.
Rogin, Leo, 319n,, 389n., 993n.
Roos, C. F., 184, 520n., 528
Rosenstein-Rodan, P. N., 49n.
Roth, H., 529
Rothkegel, W., 741n,
Roulleau, G., 586n.
Rowe, H. B., 989n.
Rowe, J. W. F., 430n., 490n., 529n., 736n.,
798n.
Rueff, Jaques, 519n.
Rural Electrification Act, 1021
Ryan, F. W., 826n.
Sarle, N., 532n.
Sauerbeck, A., 471
Sauerbeck index of prices, 462ra., 481n.
Saunders, Charles, 51 In.
Saving, 40, 75-84, 230, 587
definition of, 75
deflationary effect of, 587
Savings, forced, 112n.
Schacht, H., 722n.f 881, 900
Schmidt, C. T., 766n.
Schmoller, G. von, 228n., 229n., 564n.
Schneider, E., 57n., 476n.
Schultz, Henry, 88n., 520n.
Schulze-Gaevernitz, G. von, 376
Schumpeter, Elizabeth B., 249n., 26 In.,
459n., 469n.
Schuster, Sir Arthur, 166n.
Schuster, Sir Felix, 677n.
Schwartz, G. L., 799n., 967
Scott, R. W., 245rc., 247, 248
Scoville, J. W., 773n.
Seasonal variations, 20-21
Secondary factors, 25
Secondary reserves of banks, 644
Secondary wave, 145-150, 181
Secular trend, 203
Securities Act of 1933, 986-987
Security issues (see Issues)
Security listings, new, 599
Self-generating theories, 139n.
Self-reinforcement, 181-182
Seligman, E. R. A., 826
Semeiology, 14-24
Sensitive Price Index, 23n.
Sequence and lag, 57n.
Series (see Banking; Clean cyclical; Clean
trend; Consequential factors; Erratic;
Individual; Natural; Primary; Syn-
thetic; Systematic; Time)
Shannon, H. A., 378n.
Sherman (silver) Act, 318
INDEX
1093
Shipbuilding cycle, 533-535
Shipping, American, 383-384
German, development of, 352
and shipbuilding, British, 368-370
Siemens, W., 373
Silberling, N., 264n., 469n.
Silberling index, 469
Silver, demonetization of, in Germany, 308-
309
Silver policy of United States, 1876-1896,
317-318
Simiand, F., 32n., 565n.
Singer, H. W., 63n.
Single-cycle hypothesis, 161-164, 170
Slichter, S., 786n.
Sloan, L. H., 830n.
Sloppiness, 44, 66
Slump, of 1937-1938, 1016-1020, 1026-1032
(See also Public spending)
Slutsky effect, 180
Smart's Annals, 267n.
Smith, Adam, 15rc., 127n., 278
Smith, B. B., 23n., 813n.
Smith, W. B., 252n., 470n., 582n.
Smith, W. B. and Cole, A. H., 615n., 630
Smithian principles, 346
Snider-Persons index of sensitive prices, 525
Snyder, Carl, 23n., 24n., 31n., 35n., 427,
460, 461, 477n., 479, 489n., 490n., 523,
551n., 585-586, 590n., 593n., 598,
628ra., 745w., 797n., 837n., 1000
Social security payments, deflationary
effects of, 1032n.
Soetbeer index of prices, 462n.
Sombart, Werner, 223, 228, 229n.
South African and Australian mining boom,
1895-1896, 381
South African gold, modus operand! of, 676
South Sea Company, 248n., 249
Special banks (see Banks)
Special cycles, 178, 179, 533
Special trend, 205
Specie Circular of 1836, 293
Specific cycles, 178n.
Speculation, 54
stock, and central banks, 689-691
United States, postwar, 874-880
Speculator, definition of, 679
Spending, riddle of, 552n.
rate of, 546
(See also Government; Public)
Spiethoff, Arthur, 7n., 12n., 19, 33n., 137n.,
155n., 163n., 164, 165, 220n., 221n.,
224, 254, 281n., 351, 362, 377, 379n.,
380, 436n.f 446n., 448, 470n., 475n.,
485w., 680n., 766
Spiral (see Vicious spiral)
Sprague, O. M., 582n.
Sraffa, P., 90n.
Stackelberg, H. van, 57n., 60n.
Staehle, H., 528
Stamp, Lord, 343, 563n., 729n., 799n.
Standard Oil Company, 385, 418
Standard Statistics Company, index of
bond yields, 619n.
Standard units, 21
Standstill agreement, 93 In.
Stationary process, 35-38
Statistical normal, 206
Statistical universe, 194
Steam engine, 273-274
Steel famine, 426
Stein-Hardenberg legislation, 257
Sterling bloc, 963
Stewart, E. B., 963n.
Stewart, W. W., 490n.
Stochastic or random variables, 194-197
Stock and Produce Exchange Act, German
(1896), 436-437
Stock brokers, 679
Stock Clearing Corporation, 680
Stock dividends, 621
Stock exchange, 612n.
Stock exchanges, 678-691
Stock jobbers, 679
Stock prices, 16
and devaluation, 1000
in 1932, 945
Stock pricing and easy money, 683-684
and mass psychology, 682-683
peculiar nature of, 681
rationality of, 683, 684, 689
Stock speculation (see Speculation)
Stocks of commodities, 16
Stockungsspanne, 164
Stoker, H. M., 458n., 470n.
Stolper, W., 964n.
Stratton, H. J., 419n.
Strauss, F., and Altschul, E., 738n.
Strong, B., 899n.
Suasion, 651
Substitutability of factors, 39
1094
BUSINESS CYCLES
Supply of monetary metals, influence of,
231-233
Sweezy, M. Yaple, 1039n.
Sweezy, P. M., 56n., 64n., 107n., 245ri.,
839n.
Synchronization of production process, 40
Synthetic series, 18
System expenditure, 548-561
Systematic series, 25
Tappan-Hollond, M., 830 n.
Tariffs, American, 259, 309-310
Taussig, F. W., 3G7n., 395n., 422, 667
Taylor, A. E., 532/i.
Taylor, G. R., 458n.
Taylorization, 783n.
Tax, capital gains, 1040
payroll, 1042
undivided profits, 1040-1041
Taxes, transference, influence of, 731
income, corporation, and estate, effects
of, United States, 1039-1040
Tebbutt, A. R., 914, 928n.
Technocrats, 496n.
Technological lags, 48
Temporary investment (see Investment)
Tennessee Valley Authority Act, 988/t.
Tension (money market), 17
Terborgh, G. W., and Gray, T. II., 746n.
Teubert, W., 491n.
Textiles, 240, 242
Theoretical variables, 194-197
Thirty- Years' War, 231n., 234
Thomas, C. E., 549, 582n.-584n., 621n.,
633n., 888n.
Thomas, Dorothy S., 24n., 180
Thomas, Woodlief, 490n.
Thorp, Willard, 12n., 789n.
Thorp's Annals, 221n., 298, 327n., 362n.,
365n., 377n., 379n., 380w., 427n., 436w.,
446n.
Timber, scarcity of, 242
Time deposits (see Deposits)
Time rate, 624
Time series, 22
analysis of, 197
definition of, 197
Timoshenko, V. P., 178n.
Tinbergen, J., 55n., 179, 184n., 185n., 186-
188, 525n., 533
Tintner, G., 57n., 193n., 199n., 201 n., 204n.f
520n., 526n., 527
Tooke and Newmarch, 343n., 46 In.
Total output, concept and measurement of,
484-491
Trade agreements, reciprocal, 1026
Transference taxes, influence of, 731
Transport of miscellaneous merchandise,
17
Treasury, United States, 113
Treasury bills, 624
Trend, breaks in, 202
meanings of, 200-205
Trend analysis, common sense of, 21
Trends (see Descriptive; Real; Reference;
Result; Secular; Special)
Tripartite agreement, 1027
Trustee Acts, English (1889 and 1893),
430n.
Trustified capitalism (see Capitalism)
Tucker, A. S., 667n.
Tucker, R., 473n., 565n.
Tugan-Baranowsky, M. J., 178n., 220n.,
254, 276, 501n.
Turnover tax, 711
Tuttle, P. M., 316n.
Tyszka, Carl, von, 565n.
U
Uncertainty, 54
Uncollected items, 890n.
Underconsumption, theories of, 141
Underemployment in perfect competition,
52
Underspending, 579, 585, 587
Undivided profits tax, 1040-1041
Unemployment, 509-519
cyclical, 515
definition of, 43
in depression, 517
disturbance, 513
in England, 1931-1938, 967-968
in equilibrium, 59-60, 67, 158, 161
normal, 511
postwar, 803-806
secondary, 515
structural, 511
technological, 513
vicarious, 513
Unemployment Relief Act of 1933, 1001
INDEX
1095
Unemployment Trust Fund, 1042
Union Pacific, policy of, 406-408
Universe, statistical, 194
Unwin, G., 236
Usher, A. P., 85n., 226, 227n., 237n., 241n.f
244n., 248n., 254, 271n., 273n., 368,
458n., 459n.
U. S. Steel Corporation, 16
foundation of, 410-411
Usury, 614, 615n.
Variables (see Historical; Stochastic or
random; Theoretical)
Veblen, Thorstein, 32
Velocity, of circulation, 545-546 v
of deposits, 579, 585, 856
Verhulst formula, 492
Vibrations, 184, 219n.
Vicious spiral, 151, 181, 925
Villard, H. H., 820n., 996ra.
Vinci, Felice, 185
Viner, J., 264n., 601n.
Volkswirt, der Deutsche, 830n.
Volterra, Vito, 184n.
W
Wagemann, Ernest, lln., 593
WagenfUhr, R., 490n., 798
Wage bill, 566-574
Wage rates, 567-574
postwar, and effects of, 834-850
Wages, behavior of, 564-578
in and after crisis, 952-954, 978-988,
1009, 1015-1016
in depression and recovery, theory of,
953-954
Wald, Abraham, 21n., 46n.
Walker, Gilbert, 755n.
Wall, N. J., 796n., 926n., 929n.
Walras, Leon, 36, 46, 47, 50, 52, 53, 78n.,
129n., 183, 309, 452n., 547
Warburton, C., 820n., 823, 824n.
Wardwell, C. A. R., 165, 210n.
War effects, 701-708
Wars, recurrence of, 697n.
Warenhandels-Krisen, 257
Warren, G. F., 323
Warren and Pearson, 520n., 738n., 796n.
Warren and Pearson price index, 471, 490n.
Water power, use of, 242
Watkins, M. W., 768n.
Weaver and Edie, 585 n.
Webb-Pomerene Act, 782
Weber, Max, 228, 229n.
Weedon, W. B., 459n.
Weill, N. E., 674n.
Weintraub, D., 518n., 805n., 806n.
Wells, O. V., 532n.
Welter, E., 723n.
West, Sir E., 267n.
Westcott, D., 926n., 927n,, 966n.
Wheat program, effects of, 991
Wheatley Act, 752, 918
Whetham, C. Dampier, 321
Whitman, R. H., 528
Whitney, M., 237n.
Whitney, S. N., 217n.
Wickens, D. L., 321, 745w.
Wicksell, K., 59, 60n., 127, 456, 546n., 580,
604, 605n.
Williams, J. II., 667n.
Williams, John Burr, 41 In.
Willis, II. P., 650n.
Wilson, E. B., 166n., 199n., 474
Windfalls, 132
Wirth, M., 220n.
Wisniewski, J., 21n.
Withers, Hartley, 673
Witte, Count S., 670n.
Wittstein formula, 492n.
Wolman, L., 565
Wood, Frances, 46 In.
Working, E. J., 520n.
Working, H., 520n., 586n.
World crisis (1931-1932), 924-954
causation of and Federal Reserve policy,
902
output in the, 924-930
in United States and Germany, 911-954
World's price level, 462n.
Woytinsky, W., 473n.
Wright, Philip, 474
Wright, Sewall, 532n.
Young, A. A., 459n., 582n., 584n., 594, 595,
644n., 851n., 859n., 864n., 865n., 867n.
Young plan, 920
Zeuthen, F., 57n., 58n., 63n.
Zinn, 184, 202n., 21 9n., 633
Zollverein, German, 280n.